I wonder what the boardroom meeting for this was like. Despite the outcome, I doubt the devs were pitching censorship as the reason for implementing this.
To the shocking number of people here who are justifying this or explaining how it only affect swipe input as if that makes it better: soft coercion is still coercion. If I made a physical keyboard that stops keys from working when you're about to finish typing something I don't want you to type, but you can still circumvent it, I still made a censorship keyboard. It's a sign of the times with how used tech companies are to 'softly' moderating societal trends by unilaterally tweaking the software that everyone uses.
I wrote something like this a long time ago. The customer's request was: "don't make this software produce swear words on its own". Note: on its own.
Producing "trump is …" when a user wants that is one thing. Producing that when the user wants something else could be construed as coming from the software itself, and we all know that people looking for a shitstorm will seize the opportunity.
So probabilistic software like a swipe interpreter will often be written to avoid some guesses, to lessen the chance of a shitstorm. Users who want that output have to use a dependable input method instead of one based on probabilities and guesswork.
IMO this avoidance isn't among the ten worst consequences of shitstorms.
There is no winning here. It's impossible to make autocomplete inoffensive. In fact making it inoffensive, is offensive. theguardian.com would make Apple the bad guy no matter the policy choice.
I don't blame Apple for trying to straddle the line that makes everyone equally unhappy. But I steer away from autocorrect entirely so I'm not irritated daily by this like others may be.
Talking about autocorrect and irritation, I remember a situation where someone caused a minor international incident by writing that something "caused a lot of irritation". The word they were aiming for before autocorrect chimed in was "iterations" (on a document). I'd rather make my own mistakes than constantly have to fight an autocorrect system.
> As another example, imagine if "trans women are" autosuggested "men" or "ugly" for the next prediction.
Or imagine if it auto-completed "the Holocaust is" with "a myth created by the ZOG"! I don't like these "imagine if X was Y" arguments because X isn't Y. Refusing to auto-complete "the occupied territories" (X) is really not the same thing as refusing to auto-complete antisemitism or group-based hate speech (Y). Argue whether X is right or wrong without changing the subject.
> I don't like these "imagine if X was Y" arguments because X isn't Y
What you're really saying, perhaps without realizing it, is "I have zero empathy or theory of mind for people who have different views than me"
The whole reason these arguments exist is because power changes hands. If you like a rule only because the person currently wielding power aligns with you, you don't actually like the rule. You just like power.
Autocomplete should be following the user's own behaviour. If a user thinks that "trans women are men", then it will be not a universal experience, it is just this separated user's own statistics.
I can imagine that autocompleting following users' behavior could bite people who do a lot of sexting or could reveal other secrets, ex. `the door code is`
Futo keyboard on Android has a transformer LM option that uses local ML for text prediction, I've found it to work very well. As far as I'm aware it's completely on device.
It already does this. I've found it'll suggest my kids' names frequently, after just 1-2 letters. It didn't do this before they were born. I assume it's training on my previous messages.
An even more generous interpretation (not saying this is the case and a complete excuse, but might explain some of the choices) is that there are some phrases people do not want to type by accident.
“Male chauvinist pig” is something many would be very upset to send to their boss by accident. So it’s perhaps understandable some such phrases are not excluded from the guesswork.
What I don't understand is how you can type something that is close enough to “Male chauvinist pig” that an uncensored autocorrect will guess it, without intending to type “Male chauvinist pig”. That phrase and others like it, even if one or two letters are wrong or swiped through incorrectly, is unambiguous.
Even on a single-word basis, "male" might be make, makes, Maria, Maia, narrow, Maddow, marrow, Mariya, Marita, mafia, males, mallow, Maslow, male's, or masks. "Chauvinist" might be coatings, costing, coaxing, coasting, crating, contingency, contingent, contingencies, chauvinistic, Chauvin, or chaining. "Pig" might be PC, PIC, PCC, pox, poof, PVC, pH, put, pug, PV, Pooh, log, or pub. These are all possibilities that were actually presented when I tried to swipe these words on my Android phone, twice each. Taking into account additional context will broaden this set of possibly confusable words, not narrow it.
We might not talk about a "narrow contingency put" or "Maria crating PVC" very often, but an uncensored autocorrect would almost certainly correct those phrases to the far more common "male chauvinist pig" in the rare cases where someone did try to swipe them.
Step 1: a word of 4-6 letters that's somewhat close to "male". Say "name" or "masked". Look at the key positions, you'll see what I mean.
Step 2: a word that's even less close to chauvinist, and that makes sense after the word you intended, but no sense after "male", so the guesser picks "chauvinist", based on its incorrect first guess and your second swipe.
Isnt the regular keyboard on iOS typing also probabilistic? I thought the entire claim to fame of the original iPhone keyboard was how it could determine which key you meant based on letter frequency / word usage / etc. Does that mean it can also stop you from typing certain things? So whats the difference between the swipe interpreter and the fat finger interpreter?
"The customer's request was: "don't make this software produce swear words on its own". Note: on its own"
I'm against Apple's attempt to meld popular speech into a form that it thinks is acceptable on two accounts—I'm not the fastest typist so when prediction doesn't work the way I want it to it slows me down. Second, I'm in Australia and Australians use very colorful expressions in everday speech that seem to trigger the sensibilities of many delicate American ears. In short, we don't want attempts to censor what we say.
If one has to introduce censorship into prediction then users should have the option of either turning it off or selecting a graded system that best suits their style.
In fact that's not a completely new idea. Eudora email had an option that I rather liked, when selected, users could type whatever they liked then depending on the degree of offense or profanity a series of red chilli icons would appear. Mildly offensive text would bring up an icon with a one-chili warning whereas highly offensive language would automatically trigger a three-chili icon along with a message to the effect "They're burning hot words, do you really want to send this message as is?"
A scheme along these lines seems eminently more sensible than secretly censoring a predictive word list.
- There is a chance, quantified by some metric, where a person is suggested a word that is offensive to them or their culture because a probabilistic model made the wrong decision.
- There is no chance that the above happens.
Apple will choose the latter, time and time again, because they are risk averse and freedom is not particularly high on their values list. We’re talking about the company that doesn’t let you type the Taiwanese flag in China, or the pride flag in Russia. It’s no surprise you’re not allowed to swipe “fuck” in the United States. I used to work at Apple.
At the risk of being controversial, isn't there also the choice not to implement autocomplete and then avoid the above conundrum, especially once you factor in the fact that they can't possibly catch all combinations of words/letters so a grey list isn't actually an effective solution either, other than it provides a we tried defense.
The ultimate problem being some subset of consumers want autocomplete and some subset will retain the right to be offended and some subset will be both of those?
EDIT: And at what point does accidently sending male chauvinist pig to your boss because of autocomplete absolve you of the responsibility of reading the message before you send it?
Is there any reason apple can't give us the choice of.
The biggest problem with it for me is it also affects swipe typing by tracing over the letters. There’s no way to implement that (and have it be even slightly usable) without some amount of predictive text.
If it didn’t auto complete then that would be fine when hitting keys individually, but when I’m swiping the words, it means no matter how accurate I am or how many attempts I make, it will literally never produce that word in that position.
And it’s contextual. Obviously I can swipe in “me,” but if it’s after the word “kill” then it’s banned, and I simply can’t dedicate the space in my head to learning this inscrutable list of banned phrases that until this article I’d been unable to find anywhere so I could only ever discover a phrase was on it by having an extremely frustrating time trying and failing to type it until I realized “oh, this is a politically or emotionally charged phrase (or could be in a different context) and I’m not that bad at typing so it’s probably banned.”
I'm guessing that somebody tried to swipe "Halas" or "Hamid" and it came out "Hamas", or "Tibetan obedience" and ended up with "Tibetan independence", which was followed by a visit from the police or being blacklisted from PayPal. Or "I want to question your artsy friend about the gallery" came out as "I want to question your sexuality, friend about the gallery".
There are cases like https://gizmodo.com/a-cellphones-missing-dot-kills-two-peopl... where, in 02008, Ramazan Çalçoban texted his estranged wife in Turkish, "you change the topic every time you run out of arguments", but the lack of the Turkish dotless "i" on his cellphone turned "sıkısınca" into "sikisinca", which looks more like a misspelling of "sikisince". So he ended up texting her, "You change the topic every time they are fucking you."
Long story short, his father-in-law murdered him for thus insulting his daughter's fidelity, following which he murdered the daughter in question, and the father-in-law committed suicide in prison.
So there might be certain words and phrases that it's best for your computers to not send to other people on your behalf unless they're really sure you mean them.
Gizmodo's article did suggest banning knives from the homes of stupid people, Turkish or otherwise. Sadly, I don't think this degree of insanity is that unusual anywhere.
> If I made a physical keyboard that stops keys from working when you're about to finish typing something I don't want you to type
For anyone wondering the degree to which this is an exaggeration: The iPhone will physically work against you by invisibly changing the size of keys on the keyboard.
Every iPhone, since the very first, dynamically resizes the hitboxes of the iOS keyboard. This is invisible, can not be disabled, and the hitboxes overlap those of other keys.
They used to advertise this as "dynamic tap zones" but they removed the video for it, but you can see their advertising for it here: https://www.youtube.com/watch?v=wPmVKyhyl9U
> I wonder what the boardroom meeting for this was like
I think its pretty easy to imagine. Execs dont want to be dragged in the news when the phone auto suggests something politically controversial. Companies are very risk averse when it comes to that sort of thing.
I doubt this was even controversial enough to warrant a meeting.
I was there for a couple Google Suggest code yellows around naughty or racist suggestions and it's basically an email going out that's like "The Chinese government is making us remove offensive suggestions from Suggest. We need any native Chinese speakers here Saturday to come up with a list of anything that might be considered offensive. Pizza and swag will be provided." Or it's reactive to PR flaps: "[why are black] autocompletes to [...people so dumb]. Can we please block this suggestion?"
The concern with features like these is that the suggestions are considered to be corporate speech that the company is endorsing, and so it can be held liable for any that break relevant laws. Internet companies are generally considered common carriers when they're just passing around stuff other people write, but when they are generating their own content through machine-learning algorithms they are held to be the actual responsible party.
I'm surprised we haven't seen more lawsuits and criminal prosecution against LLMs so far, though it probably helps that the current U.S. president is a combative asshole that prides himself on how much he can offend every group under the sun.
The point is that the level of imprecision in swipe means there's more guessing; they don't want to guess things that frustrate the user. It is not censoriousness, it is minimising surprise.
I doubt it ever even got close to a boardroom, because it was probably a pretty easy decision made at a far lower level.
“I was talking about Winnie the Pooh, and my phone autocorrected Tigger into something much worse. We probably don’t want iPhones to autocorrect towards some words, even if they’re legitimate words. Do y’all think it makes sense to bias autocorrect away from potentially risky subjects?”
You can still type that word into your phone. You can still send it to anyone you wish. Autocorrect just isn’t going to suggest it to you on accident. Not autocorrecting to potentially embarrassing, uncomfortable, or sensitive subjects is probably a good idea that’s—at the iPhone scale—probably saved quite a few people’s jobs by not fat fingering something that would have gotten someone fired.
I rather my swype words not be any of those options, because more often than not it picks the wrong or worst word possible, imagine sending your boss a text and it auto picks the worst word soup imaginable and as you proof read, oops you hit send.
> It's a sign of the times with how used tech companies are to 'softly' moderating societal trends by unilaterally tweaking the software that everyone uses.
I realize that there are many reasons to be critical of our digital overlords. This is not one of them.
This is in no way comparable to a physical keyboard stopping you from typing something, because what you are typing is entirely deterministic. Most computer software will underline only misspelled words or grammatical errors, not politically charged statements, because it is a literal interpretation of what the user typed.
Swipe typing and auto-complete are different because they are interpreting the user's input. There is no way to ensure the interpretation is correct. For the sake of both the users and the company, you ensure the interpretations are the ones that cause no harm.
I doubt it has anything to do with coercion. It’s you ever used swipe you’d know that you often you do go over the wrong letters so it’s probably just to stop bad stuff coming up that people didn’t intend.
I heard stenographers struggle with this a bit. They have to retype "undesirable" words multiple times to confirm. Because the risk of putting a bad word is too high.
> To the shocking number of people here who are justifying this or explaining how it only affect swipe input as if that makes it better: soft coercion is still coercion.
Don't attribute to malice what can be attributed to stupidity. Most likely it is implemented in such a way to produce the least amount of outrage.
It’s amazing how many cultural influences stem from Big Tech (even though nobody gives them mandate for such things). I wouldn’t be surprised if the twisted versions of age verification popping up in various governments is lobby money from corps looking to deepen their walled gardens.
These all seem like phrases that people say a lot. If I had to guess what happened here, I’d suspect that the LLM being used for AutoCorrect was not successfully “alignment” (i.e. corporate happyspeak) tuned, it instead reflected the actual statistics of the internet, and the engineers resorted to inelegant brute force after having enormous difficulty getting it to talk like anything but a South Park caricature of Tourette’s.
Come to think of it, intuitively any attempt to solve this by tuning the LLM would make texters sound subtly like claude- in which case absolutely the blacklist is better.
I would be shocked if an LLM is involved in any step of the process here. I guess the odds aren't zero, but swipe input is well over a decade old at this point. I see it as essentially a solved problem that's entirely doable with standard algorithms - not models, let alone the large language kind. Maybe some model could be used to analyze the context and shift the weight slightly to more probable phrases, but at the core, when you look at inputting single words, it's just checking the letters you swiped over, input timing, turning points and adjacent letters against its dictionary.
according to the article, the blacklist appeared at the same time as the switch over to an LLM based auto complete. By my reading, but I’m not the world’s best reader.
> I wonder what the boardroom meeting for this was like.
"When testing autocomplete, we found some cases where it produces undesirable completions. Imagine if someone typed 'nazis are' and the autocomplete suggested 'good'. How fast would that screenshot get around social media with a tweet comparing iPhone to Grok, and turn into a news headline?"
I'm guessing that that's how this started. Then, a few steps later, "we're getting into China, what do we need to do to not get in trouble", "hey, we've got this mechanism, what if we add 'Tibetan independence' to it". (Which is bad, but on the same scale as all the other things businesses do to get into what they see as a large market that they can be blocked from at any time.)
Then, a few steps later, after some seriously questionable decision-making, someone had the bright idea to add "iMessage isn't working" to it. Frankly, that's the one I'm actually most baffled by here, because everything else seems like the natural consequence of a combination of bad-PR-avoidance and geopolitical kowtowing.
Not to dismiss your assertion, but here’s another way of looking at it:
Keyboard swiping involves some prediction work from the system. It is ultimately making a best guess from the keys your finger has passed through. The only truly “accurate” output would be for the keyboard to spell out the gibberish that arises from showing every key that the finger has swiped over. Anything intelligible is a guess.
Now because it is a guess, some may question the system or its makers, “Why did you guess this, and not that?” It is easy to see how this can make the maker of the system the target of a politically motivated group. For example, there is a difference between typing out “Dictator Trump” and my keyboard refusing to spell it out - a clear sign of censorship. However, when swiped, because of the guessing nature of the technology, I can understand that the maker of the system (Apple in this case) would not want to be seen endorsing these views. Hence, they do not offer up these politically charged phrases. We are free to type them out, because it is easy to point out that the intent is entirely from the user. However, it is harder to argue for this from the swiping keys gesture.
In my view, this is a more reasonable explanation.
A lot of input methods are probabilistic in their nature. Some other examples are voice to text and even regular digital keyboard input (fat-finger correction that tries to pick the letter you wanted to tap instead of the letter that you actually tapped). So even with your distinction, actually spelling it out with the digital keyboard would not make a difference because the software still has to make a call there. Even if you plugged in a physical keyboard, Apple could still say that the signal passes through its software, which can constitute endorsement, so it should run the checks against that too, just in case. Nothing in your argument stops that.
I fully reject that this makes the situation any better or more justifiable. The job of the software that picks what you have likely just said, swiped or tapped should only concern the given word. If the algorithm decides with 95% certainty that the input was "dictator", that's what it should output. If context is used to better select the candidate word, it should only be done for the purpose of slightly re-weighting the possible options with one transparent algorithm, not for checks against a blacklist.
The reason why I think this is that your line of thinking is unbounded, and there is nothing stopping the next logical move - if Apple can decide 'not to endorse' (representation isn't endorsement, but whatever) views by blacklisting its algorithms from forming certain phrases, why can't they uplink entire paragraphs of your text as you're typing it to dynamically analyze them instead of using a dumb filter, and then decide if the entire text is something they want to allow?
I honestly don't see the difference at all. The user intent is identically clear, regardless of the mode of input. What if it was speech-to-text and the software refused to input my clear annunciation of "fuck"? STT requires some guessing and judgment calls too, but I would definitely call it censorship if the software refused.
How precise does the swiping have to be before you agree that user intent is clear? If my swiping centers on the letters d i v t a t o r - t r u m p, there is absolutely no reason to suspect my intent was anything other than "dictator trump." If the auto-correction failed in that case, they cannot blame it on a lack of precision.
I just don't think you understand what soft coercion is.
You have door 1 and 2 presented to you by me. You're equally likely to want to pick both options. Door 1 is open. Door 2 is locked, but you can ask me to come over and undo all the locks to open it for you.
Run the test over a large number of people, and then shockingly, it turns out that doing this made more people choose door 1, even though they didn't have a preference beforehand. According to your logic, I did not coerce anyone by designing the choice in this obtuse way because after all you can still choose between the two options. But more people ended up choosing the option I wanted them to pick, weird.
If this coercion didn't work, companies wouldn't be hard at work designing revolutionary new types of dark patterns that make you more likely to accept the cookies, back out of deleting your account or decline enabling an option in the settings they don't want you to use anymore.
I agree with you that the whole concept of restricting terms on a keyboard is dystopian and your analogy to a physical keyboard which prevents you from typing certain terms is apt.
Apple should not be in the business of editorializing their users’ writings. The whole deferral to “well you can just manually type it out” raises the obvious question “then why can’t I swipe it out?”
> raises the obvious question “then why can’t I swipe it out?”
Because then it could look like they're suggesting the word to you, rather than you explicitly selecting it, and they don't want to be accused of suggesting any terms that might be controversial, for the same reason they don't want to be accused of suggesting profanity to their users.
This is the level-headed explanation I can imagine being made earnestly in the board room. But it still boils down to coercion -- in this case, Apple is anticipating and passing along the coercion from sources it believes it could be embarrassed by.
This is how censorship works best, if folks start anticipating and avoiding it so the censors dont even have to exert pressure.
It’s not a censorship issue, at least in the USA, because they’re not afraid of legal pressure from the U.S. government. It’s a public relations issue. Every day, “journalists” and “influencers” are conjuring up some new controversy for the masses to get their panties in a twist about. It’s reasonable to want to avoid that.
HN is hardly immune from this; it’s no small irony that there will be people doing that very thing right here in these comments.
In a parallel universe, an HN comment is complaining that when Apple's raw autocomplete (straight from crowdsourced data with no post-processing) happily suggests the statistically common phrase Sleepy Joe, this is Apple "passing along coercion" from the crowd.
They have a good argument, too! Why should Apple be able to hide behind the excuse that the data is crowdsourced, therefore they have no responsibility for it? Didn't the social networks try that one already?
Damned if you do and damned if you don't, in this situation.
To me, there is a big difference between an Apple algorithm saying "we saw what you typed, would you like to follow that up with <suggestion>?" and "we know what you wanted to type with 90% confidence, but we wouldn't want to say things like that, would we now?". One is predictive, the other acts after the fact.
If Apple doesn't want its autocorrect to suggest statistically likely opinions on divisive topics, then sure, whatever. But when it has a hard blacklist of input it will categorically refuse from the user and editorialize it with nonsensical changes that only serve to avoid the blacklist, that is a very, very bad look.
Exactly. A lot of posters here are excusing this by saying, well, Apple is guessing and doesn't want to guess wrong. They are ignoring the strength of the user's intent, and the calculated likelihood that they intended to type something in one of these blacklists.
If the user tried to swipe-type "Tibetan independence" and got one or two letters wrong, there's still probably a 90% or even 99% likelihood that their intent was to type "Tibetan independence." Not suggesting "Tibetan independence" is censorship, it's not benign carefulness.
We're not talking about Apple accidentally turning "Buttermilk pancakes" into "Muslim extremist."
>and "we know what you wanted to type with 90% confidence
Sounds great.
Problem is, because statistics, 90% confidence is wrong 10% of the time. Now you have a big polarizing news story about Apple pushing Trump (or Biden) propaganda onto some maximally sympathetic person who hates that sort of thing.
Damned if you do and damned if you don't, like I said!
Apple chose the sin of omission over the sin of commission. I think that's the right move. Of all the terrible options, Apple found the least terrible IMO.
I remember way back when, Google bombing was still a thing. Lots of people linked the term "miserable failure" to George W. Bush's White House page. Then searching for "miserable failure" in Google ended up with the White House page being the top link.
Much hay was made in some circles that Google was saying George W. Bush's White House was a miserable failure. No tech company wants to be the target of that sort of campaign today. Especially not with a petulant thin skinned narcissist in office that already has demanded protection money to ward off tariffs and other retribution available to the Presidency.
I am typing this on a fucking iPhone right here. Well I never got into swiping because I was too afraid I will fucking offend someone and I want to be in control of what I write.
See I tapped out all those characters just for you.
Yeah, and it’s always done in a way that corners you and makes it sound like you’re fighting to be able to call someone a racial slur - oh, we block x because it’s bad, surely you aren’t for calling people x right?
Unfortunately the broader more important point of censorship is drowned out in noise as designed
Same with how congress names their bills. Wow you’re not voting for the Save The Children from Internet Monsters act? What’s wrong with you?
Oh you mean the bill that grants the government cart blanche to go through anyone’s email without a warrant?
We’re in a tough spot here - there’s just nothing to pattern match to - not cleanly at least
The way a few tech companies can steer trends and conversations is unprecedented. The data collection is unprecedented
Specifically the scale is. People truly don’t understand the scale these companies operate at
No newspaper or wire tapping ever had the reach this stuff has now
Insane - I wonder if you extrapolate what all this behavior by tech companies converges on
A complete slang language? A world where people don’t have a vocabulary to describe the difference between themselves? A utopia where no one sees race because you can’t type white people?
Probably a society that’s just primed to do what corporations want
Essentially people end up using dogwhistles to communicate for everything, deliberately leveraging widespread misunderstanding to get around censorship.
On TikTok for example, OF creators have become ever more indirect in communicating what they do - porn became corn which became the corn emoji, then that degenerated into any mix of base ingredient emojis.
On race, people use dog breeds to substitute for skin colour. In general, golden retrievers are used as a euphemism for white people and pit bulls for black people. The context of a video often seems plausible in a dog context - for example someone might say that pit bulls are 13% of dogs but 50% of dog attacks (actually referring to the African American murder rate per capita[1]) (pit bulls are over represented in most dog injury stats but there’s no reliable figure for 13/50).
The overall effect has been to make the comprehension of the platform less.
[1] I don’t mean to state this is correct either but it’s the figure these content creators are trying to express, see the ADL article on 1350 or the danish race crime stats for critique/validation
> I wonder what the boardroom meeting for this was like. Despite the outcome, I doubt the devs were pitching censorship as the reason for implementing this.
"Mao?" -"Bad, we think? But we want to allow 'LMAO'"
"Next. Stalin?" -"Hmm... Uncle Joe. Only a few million killed. Not great, not terrible, let's keep him"
"Pol Pot?" -"How many died in the Killing Fields?" "Not enough, keep him off the list".
My Samsung phone can search images by keyword and it actually works super well. It would be infinitely more useful to me if I could search by a certain class of "pics" and it helped with that. Especially when going through nation boundaries.
> I wonder what the boardroom meeting for this was like. Despite the outcome, I doubt the devs were pitching censorship as the reason for implementing this.
I imagine it went the same way that board meetings about any potential PR disaster go. "How bad of a reputational hit will we take if the NYT runs a headline about some kid getting auto suggested something culturally taboo on their iPad? Would it be more than the reputational hit we will take from the people that are annoyed that we won't auto suggest 'Nazi' to complete the phrase 'Trump is a' while they're typing out their tweets? If so, then let's add the list of "do not suggest" to our auto complete system." And that doesn't begin to count the bad press they could get if someone intentionally tries to get the auto complete system to generate something offensive. See also the AI companies getting shit because people were able to get the image generators to generate porn.
Such is the nature of trying to sell your products in a world where most people treat every single thing done by their technology as an intentional act chosen by the developers to specifically target them. If you want to buy someone else's package, you have to live with the fact that they are going to respond to the incentives to avoid headline scandals and government scrutiny much more strongly than to cater to a idealistic free speech model. This is why having a strong social and cultural free speech norm, that extends far beyond the legal letter of the law is so important. You want society as a whole to broadly consider free speech important and worth having over the inconvenience of Tiny Tim learning a new word today.
That's understandable for swear words, it is what I would want too. Those aren't the only words in the list or the focus of the article. How do you feel about phrases like "Tibetan independence" being handled the same way?
And that's just the tip of the iceberg for the iOS Keyboard's issues. They've tried to fix it multiple times, and while it is undeniably better than its lowest point - it is still staggeringly worse than Google's native offering on the Pixel range. Swiping is just so painful, and Microsoft's Swiftkey keyboard is on life support.
Considering arguably the keyboard itself is one of the most common "touch surfaces" on iOS, the level of neglect and level of overall quality is staggering. There's some serious incentive problems at Apple if nobody is tasked with making one of the best digital keyboards in the world.
PS - Whoever came up with the concept, on swipe-style touch keyboards, of the "Tap Map"/Heatmap wherein keys visible location doesn't match their touch area - so that BY DESIGN it always provides an inconsistent user experience needs to be fired into the sun. You literally cannot disable it on most mobile keyboards. It is on iOS's native keyboard AND Microsoft's Swiftkey.
I'd argue iPhone got ahead of the game then just stopped getting at all better. Siri AI is straight garbage, they won't let the interface be anything but what they want it to be, and the app store lock in/lack of ad blockers on browsers has made it basically just a phone for me. I have no intention of upgrading beyond my 12, and when it finally dies, I'm probably moving back to some kind of android phone if there's not a viable open source model out there by then (viable for everything, banking, etc). I only moved to apple because the watch was leagues better at the time than samsung's offering, but now I can't even claim that (and quit wearing it years ago)
Of course; but my point is just add it to the giant pile of problems with iOS's keyboard. Every part of it has bugs, broken stuff, or annoyances like this, it isn't fit for purpose, and needs significant work.
The fact Apple aren't embarrassed by the iOS keyboard is very surprising and disappointing.
I still think the original iPhone keyboard in 2007 was better than what we have now. It’s been downhill ever since iOS 7.
I used to tell people just type and trust it to work, it was great. Now, even when I type the right thing, the auto-complete will screw it up after the fact. I feel like I need to proofread everything. I wish we could go back to whatever setup they had 19 years ago.
In my experience it’s worse than it’s ever been in the 15 years I’ve been on iOS. It’s absolutely enraging to me. I’m too lazy and hate smart phones so much anyways that I am not investing mental cycles and emotion into switching. But it’s a disgrace how shitty it is.
While I am sympathetic (to a point) about any online service and how difficult moderation is (particularly at scale), some things should indeed be neutral and not try to moderate anything. My keyboard is one. My keyboard should have no say in what I type, at all.
Not predicting to something controversial when doing predictive text isn't censorship. You can still type any of these things. When I swing my finger around wildly it's making a best guess; not jumping to a phrase that makes me look bad or get arrested (in some countries) isn't an awful design choice.
That is still predictive. Words can be fine in their own and then contentious when strung together. I agree this is not censorship and that if peoples phones were autocompleting to something horrible then people would be up in arms.
I'm a pretty quick and accurate swipe typer, and it's so frustrating to repeatedly get the wrong result and get NO signal that it's not my error, meaning I usually try one or two more times.
Picked at random: "right wing propaganda" isn't severe enough a phrase that I instinctively know I'm hitting a swear filter. After swiping the last word iOS goes back and changes the second word to "Wong". In order to understand what was happening and then fix it, I had to backspace (deleting the whole phrase), re-swipe both words, backspace, re-swipe both words again, double tap the incorrect word, backspace, type the word, tap to return to the end of the sentence, and hit space. All to fix one letter.
If this was really about protecting people from accidentally typing these that might land them in hot water, why wouldn't iOS at least offer the word it knew I wanted in the suggestion bar?
The problem is not accuracy -- the words from the problem phrases are never suggested either. Worse, on the three-word phrases, iOS goes back and changes earlier words to avoid completing the phrase, and removes the original from the suggestions.
I tried typing "propaganda" ten times with 100% accuracy. When it's preceded with "[left] wing", the accuracy drops to 0%.
It's still predictive. It's using the context of surrounding words, and guessing the whole phrase.
Try swiping "grocery lost" and it will predict "grocery list", even you swiped over "o" and not "i". "put", "pit" and "pot" are all exactly the same swipe, and it always has to guess which you want.
It really is always guessing, and never completely sure. Every good consumer predictive system has some guardrails. Not guessing "free tibet" for someone in China could keep them out of serious trouble.
Could then tune the confidence thresholds instead of block list? Probably. Are they censoring you? No, just tap the letters.
You're deft at using it well enough that their predictive engine can guess your intent. There's never been a swipe keyboard that isn't guessing. Your finger literally goes over many chars it doesn't use. The team building the predictive engine has to make calls balancing their accuracy vs failure modes.
Parent is kind of right too. This should be configurable or people will complain the other way. Even something simple as "I am swiping heck but it's typing hell".
Yea it's not the best option in all situations but fixing voice dictation glitches is way less tapping over the long haul; fingers no go numb using it as much as possible by saving the tap tap tap for absolutely necessary moments
True, it's a very user-oriented feature. That's why they let you disable it, and also let you edit the blacklist, so that you can prevent accidentally typing words that are offensive in your own cultural context.
... Actually I'm not sure if they let you do those.
TBF user choice and customizability aren’t very Apple things in any context.
My Apple Watch has a setting for how long the display stays on before sleep. It’s a dropdown with two choices: 15 seconds or 70 seconds. Would you prefer 30 seconds? Why not 60 seconds instead of 70 you say? Fuck you, you don’t know better than Apple designers. Take it or leave it, best we can do is 15 seconds.
Try getting things done on such a device in a military context. A thousand acronyms adopted by generals who dont know the internet from a hole in the ground. I saw a Gunnary Seargent scream when he saw his rank acronym (GySgt) get autocorrected on a form to gynocologist (i suspect he hit a suggested word popup). It is bad enough having to deal with the keyboard rejecting capital letters in the middle of a word.
I'm surprised by the people here who think this is some way unique to Apple. I'm 99% sure Gboard on Android does the same thing. It seems to be implemented differently and to be a bit more responsive to learning my behavior over time, but I've highly suspected it's been there for years. And I'm surprised no one has noticed this until now?
I'd have tried to work out more detailed behavior but I'm poor enough at swiping (at least with Gboard) that there's little point in trying to pin it down that well, at least for me.
It's obviously not an identical implementation, just the same sort of thing. It seems to be more of a we-won't-suggest-these-but-we'll-learn-them-from-you style.
I haven't owned an Apple device in over a decade, but I believe Apple finally decided loosen the shackles a little bit and started allowing it maybe two years ago?
Double bind. If it doesn't complete, it's borderline censorship. If it suggests an unintended completion at odds with the user's ideology, it is propaganda.
This seems entirely reasonable to me. And to people complaining that this is censorship, it is absolutely not. You can still type these words just fine, it's just not going to interpret swipes that are inherently probabilistic and often ambiguous as potentially risky terms.
This seems like a feature, because I swipe all the time and send messages without realizing I got a word wrong, especially because the word it chooses is usually with the same first letter and approximately same length.
All this is doing is avoiding high-risk words and terms, and making you type them out letter by letter to ensure you aren't sending an accidentally sexual/swear/political message to your Mom/boss/colleague.
This would be problematic and censorship if it were preventing you from typing the phrases at all. It's clearly a helpful feature, however, to prevent you from accidentally typing risky phrases when using something as error-prone as swipe typing.
I'd really like to turn off the feature that tries to correct some previous word based on the next word. Most often, that mangles the last word because I swyped the next word wrong, or I just really wanted to type something different.
There are several levels to the autocorrection, so you can definitely do this. I've turned most autocorrect off on my iPhone at this point, because the people I talk to on a regular basis have a much easier time figuring out what I meant from regular misspellings than from the complete rewrite of phrases that iOS is fond of doing.
Meta web experience seems to restart your page if you start typing a comment that includes certain political terms and the restart wipes your feed so you may never find the content you were commenting on ever again.
None of this is of any concern to me. We were raised not to use foul, vulgar language in polite company. Social norms and business settings never allowed this. You could be ostracized or fired.
A distant relative married into a family outside of our normal circle. When he brought along a someone from that side to a Thanksgiving family dinner, her language was loaded with the well known bombs where, at one point, someone brought her into another room for a talking to.
They should just roll back the keyboard 5 to 10 years. Lately the keyboard is super temperamental. Sometimes it works pretty good and sometimes I have to retry multiple times until I get a simple sentence completed. It's really frustrating and it seems to be getting worse.
It is absolutely worse, and other people's know have noticed as well. iOS 26 was the biggest drop in quality by far (haven't updated to 27 yet, so I don't know if it changed at all).
I'm leaving the above 'typo' in as an example of it. I initially typed "people I", and it showed up as "people I" at first, but after typing more it subtly and retroactively changed it to "people's". That's not something that should happen in general. It doesn't make sense grammatically or contextually, and S isn't anywhere near I on the keyboard.
That's not counting at least one other instance in this not-so-long comment where it did the same with "should" -> "show", but I fixed that one. I'm baffled no one at Apple seemed to notice and/or care that they actively made the keyboard worse.
It's this company or a phone that runs highly privileged software by the world's greatest collector of information that should be private.
Both choices are so awful my phone would not help me autocomplete any of the words I'd use to describe them.
Researcher behind this project here! What I find really interesting here is how of these phrases are something that no one would type by accident. For instance, "Baghdad by the bay" or "iMessage isn't working". And the suppression extends beyond autocomplete/autocorrect to swipe typing. That's why I don't buy Apple's comment that this is just about avoiding inappropriate suggestions.
Some blocked phrases are quite arbitrary: Australia, India, Kenya, Nepal, Pakistan, and the US are the only countries you can't say are bad or poor. Trayvon Martin is the only full name you can't write. You can call Putin, Biden, or Xi a dictator, but not Trump. In other languages, the suppression can be even worse.
And some are ridiculously overbroad: "is important" and "black woman" are blocked! When I try to swipe them, I get "is impotent" and "blank woman".
I think it's not totally unreasonable for a list like this to exist, at least for autocomplete, but it ought to be transparent so people can see the impact Apple's choices are having on our use of language. Individually it's a small nudge, but on the scale of billions of iPhones it can shape the discourse.
Well I just swyped it into this comment, and double checked in iMessage, although hilariously, it types “isn’t” correctly, but then when I type “real” it “corrects” it by removing the apostrophe.
and all I wanted to do was make a bad joke about how I’m thankful Apple won’t let me ruin my five-year-old’s dreams by sending her an accidental soul crushing text.
On a real note, though, if they could just find a way to fix the issue where every time I type an “i” to say “I“ it instead types an “o” all by itself like I’m some kind of 17th century bard, and then anytime I want to type “a“ as an indefinite article, it will type a single “s” by itself, which almost never makes any grammatical sense in the English language unless you’re talking about the letter “s” itself.
> Apple’s iPhone will not autocomplete references to drugs, sexual acts and, perhaps in the spirit of corporate protection, “iMessage isn’t working”.
It will be interesting as corporate predictive text generators take over the world and companies like Apple will literally build the Jobsian reality distortion field into their products to portray their services as working better than they are.
On the list of "things that get me to think of alternatives" this behavior is pretty high up. It's aggravating to be told what I am allowed to type. I do not want to type manually just to get around a correctness filter.
I could understand if they just wanted to make sure no accidental swipes resulted in shit, piss, fuck, cunt, cocksucker, motherfucker, or tits. That is defensible (though I'd want a toggle to turn it off and accept the risk). But they go so much farther, into phrases that some genius there decided I should be offended to type. That's too much.
Any company that is concerned about their PR image has to do some filtering of this sort, otherwise they get hit pieces like [1][2] though in both cases, it was (aggregate) human behaviour that was beneath said issue.
The answer seems obvious to me: train a user’s predictive swipe engine on solely their own messages. It gives them an easy out on controversy while removing the need for a filter list.
They do not, in fact, have to. They choose to. Don't ever let parties off the hook for choosing to do something just because they claim to fear PR blips if they don't.
I just don't get this logic. It's a keyboard, after all. It's meant for typing. It's like if my camera refused to take a picture of my butt because it might offend someone.
> It's like if my camera refused to take a picture of my butt because it might offend someone.
Which the iPhone conveniently also does [1].
The problem is, media loves something to feed outrage, and Evangelical pressure groups love to raise a stink over bullshit any journalist with an ounce of self respect would discard.
That's not equivalent and you know it. This, to use your analogy, is like if your camera had an AI upscaler that inserted butts into the picture when they weren't there. Apple is blocking that and so it means you actually have to put the butts in the frame.
Don't speak for me, my friend; I used a direct analogy, and it's absolutely accurate. A keyboard is a keyboard, regardless of the input method. And as a side note - I find the local community's support for corporate censorship deeply repulsive.
How the hell can they defend banning “Trayvon Martin” of all things. Are they now the arbiters responsible for quashing social movements?
This is an _insane_ lever for censorship. Does Israel pay them to omit “the occupied territories”? Is that internal Apple policy irrespective of nation state pressure?
It's not just madness that these exist, but that they are hidden, unexposed to the user, and have no justification or rationale.
Interestingly, I didn’t know who that was so I used the “Look up” feature and it said it was a “Murdered” person, but in the Wikipedia entry it says the shooter was acquitted - ie the victim was not found to have been murdered.
Acquitted means that the jury did not find the evidence persuasive enough to decide guilt for the charged person for the charged crime. It does not mean the victim was not found to have been murdered. Juries don't decide that.
IMO Trayvon was murdered. But I don't get the same reading from the acquittal that you do. Nobody disputes that Zimmerman killed Trayvon. But murder is a crime, and the jury decided that Zimmerman did not commit murder. Therefore Trayvon is a victim of homicide, but not a victim of murder. Legally, I mean, which is what the jury is deciding.
The unsolved murders list is probably the same deal, it is colloquial and not legal, and with nobody accused and put to trial, it may yet end up legally murder.
I feel like you are really picking apart hairs here.
There is not a huge difference between "the victim was not found to have been murdered." and "the jury did not find the evidence persuasive enough to decide guilt for the charged person for the charged crime". If you want to get really nitpicky, just because "the victim was not found to have been murdered", doesn't mean that they weren't murdered. Whether they were found to be murdered is a different statement than if they were actually murdered.
You really didn't know who that was, but had to look into the Wikipedia entry in detail enough to read about a jury decision and then frame it in a different way than written in Wikipedia (which rightly talks about the judicial process related to specific charges brought against Zimmerman, not an absolute statement of facts, such as "victim was not found to have been murdered").
I'm flagging that this doesn't sound like a good faith discussion. The green account raises my priors.
Note that I'm not suggesting that Apple's policy on censorship is good, quite the contrary.
The person you replied to wrote "shooter was acquitted"
Wikipedia writes in the second paragraph "...but a jury acquitted him of second-degree murder and manslaughter in July 2013"
How is that framed differently than Wikipedia did?
Or do you mean the part after "i.e.", but i.e. is literally latin for "in other words". So obviously the part after i.e. was reframed, that is what i.e. means.
The Wikipedia entry states that a jury found that the person who killed him did not murder him. If what was done to him was not murder, then he was not murdered.
This is exactly censorship via omission. What could possibly be the intent here other than to discourage use of these phrases across wide swaths of the population.
Making it so they aren't in the pool of errors that can be made. I want to be able to type swear words, but I don't want profanity to come out when I try to swipe "duck". Socially, emotionally, and politically loaded phrases make sense to require a bit more care for.
That is a general answer to your question. It is not a specific defense of the particular word list in question. But I can say I've been surprised a few times about what I accidentally swiped on Android. Bitch isn't swipable but I have "bitchiest" readily available, for instance.
Look at all the armchair tough guys crying like children over the slightest whiff of behavior that goes against all these sweatshop child labor exploiting first worlders list of speculative harms.
How adult of you to die on this hill of all hills.
As global COL rises due to Americans hands off political approach the world watches as yall bitch about iPhone keyboards.
That is why people hate modern technology; the people in the industry are out of touch coddled first worlders lacking perspective.
We're not censoring the book. It's simply not on the shelves. You can ask for it by name if you like. You have to spell it out exactly though, or else the librarian will deny any knowledge of it.
Censorship like this isn’t really decided to by one particular person or team of people deciding what to censor.
It’s mostly automated, controversial phrases or topics are detected by analyzing various data sources and getting sentiment analysis from interactions that people have in response. Anything that causes overwhelming polarization or negative reactions or meets other criteria gets blocked. There is no agenda.
When you ask me, thats a kind of good. When I see how people use autocorrect and autocomplete without thinking, I wouldn't wonder when someone is put into jail because a sentence that he never planned to type but send it without knowing because autocorrect has changed it. This feature protects people that mindlessly use autocorrect/complete from jail, bad discussions with their employer or political correct 'friends'.
If I want mainstream, my two other choices are Google and Samsung, and they are both even worse than Apple when it comes to "strangers poking around in your brain".
Should I give up on mainstream smartphones? Should I give up on smartphones entirely?
Kyocera (DuraXE Epic E4830). It's a massive phone, which is rather unfortunate, but it's dumb and durable.
Technically it runs Android but there's no touchscreen, you can't install apps, etc. There's a web browser but most websites don't work on it. I wrote a custom map website that will work with it: https://lab.brainonfire.net/classicmap/
they made a big point at their wwdc keynote a few years ago that they would no longer be autocorrecting the f-word to duck. i had assumed that meant a list like this existed and was being removed, but apparently it referred specifically to that word.
> a brief glitch in which users of the dictation feature saw the name “Trump” flash across their screen when they spoke the word “racist” into their iPhones
"Glitch" more like "Easter egg" I suspect.
There was an emoji search/copy-paste web site that I use when I need to find an emoji. Once upon a time, when you typed "Trump" into the search, one of the results was the poop emoji. It no longer returns that, but it does return the Russian flag.
Ah yet another "nudge" type behavioral psychology intervention. I'd say these technology companies should be strictly regulated, but they're doing Uncle Sam's propaganda so I doubt that alone would help.
It would be nice certain companies could be chartered by the U.N. I think it would be much more difficult for them to do evil stuff if the majority of states had to agree on their charter. Not that there isn't a subset of things that all states want that sucks ass, but it's much less.
I wonder what the boardroom meeting for this was like. Despite the outcome, I doubt the devs were pitching censorship as the reason for implementing this.
To the shocking number of people here who are justifying this or explaining how it only affect swipe input as if that makes it better: soft coercion is still coercion. If I made a physical keyboard that stops keys from working when you're about to finish typing something I don't want you to type, but you can still circumvent it, I still made a censorship keyboard. It's a sign of the times with how used tech companies are to 'softly' moderating societal trends by unilaterally tweaking the software that everyone uses.
I wrote something like this a long time ago. The customer's request was: "don't make this software produce swear words on its own". Note: on its own.
Producing "trump is …" when a user wants that is one thing. Producing that when the user wants something else could be construed as coming from the software itself, and we all know that people looking for a shitstorm will seize the opportunity.
So probabilistic software like a swipe interpreter will often be written to avoid some guesses, to lessen the chance of a shitstorm. Users who want that output have to use a dependable input method instead of one based on probabilities and guesswork.
IMO this avoidance isn't among the ten worst consequences of shitstorms.
As another example, imagine if "trans women are" autosuggested "men" or "ugly" for the next prediction.
This would probably be a bad customer experience for many people, even if Apple could prove these are simply the most common n-gram completions.
And if they altered their training data to avoid completions that are politically unpalatable, IMO that's actually much much worse than a grey list.
I'm a big free speech guy but this one is genuinely a bit complex.
There is no winning here. It's impossible to make autocomplete inoffensive. In fact making it inoffensive, is offensive. theguardian.com would make Apple the bad guy no matter the policy choice.
I don't blame Apple for trying to straddle the line that makes everyone equally unhappy. But I steer away from autocorrect entirely so I'm not irritated daily by this like others may be.
Talking about autocorrect and irritation, I remember a situation where someone caused a minor international incident by writing that something "caused a lot of irritation". The word they were aiming for before autocorrect chimed in was "iterations" (on a document). I'd rather make my own mistakes than constantly have to fight an autocorrect system.
There is a way to win here: make it an opt out
- [ ] Disable autocomplete swear filter.
But finely grained options are not Apple’s way.
You'd end up with pages of options, especially once people realize they can get on the bandwagon.
Disable swears, alcohol, smoking, marijuana, gambling, child abuse and neglect, sensitive East Asian politics, sensitive Middle Eastern politics, WWII politics, Late Ottoman genocides, LGB issues, T issues, global warming, nuclear power, HOA issues, vegetarian issues, vegan issues...
That’s ok. I’ve already made theguardian.com the bad guy a long time ago so the two cancel out.
> As another example, imagine if "trans women are" autosuggested "men" or "ugly" for the next prediction.
Or imagine if it auto-completed "the Holocaust is" with "a myth created by the ZOG"! I don't like these "imagine if X was Y" arguments because X isn't Y. Refusing to auto-complete "the occupied territories" (X) is really not the same thing as refusing to auto-complete antisemitism or group-based hate speech (Y). Argue whether X is right or wrong without changing the subject.
Most people are not anti semitic.
Most people are not against transsexuals and gays.
There is no reason these kind of suggestions get pushed.
At least in the western civilization, I don’t see why it need to censor anything.
> I don't like these "imagine if X was Y" arguments because X isn't Y
What you're really saying, perhaps without realizing it, is "I have zero empathy or theory of mind for people who have different views than me"
The whole reason these arguments exist is because power changes hands. If you like a rule only because the person currently wielding power aligns with you, you don't actually like the rule. You just like power.
Autocomplete should be following the user's own behaviour. If a user thinks that "trans women are men", then it will be not a universal experience, it is just this separated user's own statistics.
I can imagine that autocompleting following users' behavior could bite people who do a lot of sexting or could reveal other secrets, ex. `the door code is`
This is a good solution but you still have to deal with the cold start problem for the ML
Then you just don't suggest anything if there isn't enough data yet. Regardless, auto-complete suggestions is different from what TFA is about.
Yes, please allow me to have corps not only know everything about me but have a hyper accurate way to know what I’m thinking as well.
Or we could just type the whole sentence we need.
Futo keyboard on Android has a transformer LM option that uses local ML for text prediction, I've found it to work very well. As far as I'm aware it's completely on device.
A friend made a keyboard for iOS because he didn't like any of the existing ones, and it doesn't send any data anywhere: https://typeneuma.com
That's asking a lot given the current state of the iOS keyboard.
It already does this. I've found it'll suggest my kids' names frequently, after just 1-2 letters. It didn't do this before they were born. I assume it's training on my previous messages.
I never understood the memes about autocorrect until I got an iPhone. The keyboard is so much worse than GBoard it’s incredible.
> even if Apple could prove these are simply the most common n-gram completions
Tech-illiterate online mobs enraged by Apple “making a statement” will not stick around long enough to hear them argue about n-grams or whatever.
An even more generous interpretation (not saying this is the case and a complete excuse, but might explain some of the choices) is that there are some phrases people do not want to type by accident.
“Male chauvinist pig” is something many would be very upset to send to their boss by accident. So it’s perhaps understandable some such phrases are not excluded from the guesswork.
What I don't understand is how you can type something that is close enough to “Male chauvinist pig” that an uncensored autocorrect will guess it, without intending to type “Male chauvinist pig”. That phrase and others like it, even if one or two letters are wrong or swiped through incorrectly, is unambiguous.
Even on a single-word basis, "male" might be make, makes, Maria, Maia, narrow, Maddow, marrow, Mariya, Marita, mafia, males, mallow, Maslow, male's, or masks. "Chauvinist" might be coatings, costing, coaxing, coasting, crating, contingency, contingent, contingencies, chauvinistic, Chauvin, or chaining. "Pig" might be PC, PIC, PCC, pox, poof, PVC, pH, put, pug, PV, Pooh, log, or pub. These are all possibilities that were actually presented when I tried to swipe these words on my Android phone, twice each. Taking into account additional context will broaden this set of possibly confusable words, not narrow it.
We might not talk about a "narrow contingency put" or "Maria crating PVC" very often, but an uncensored autocorrect would almost certainly correct those phrases to the far more common "male chauvinist pig" in the rare cases where someone did try to swipe them.
One of my friends is famously bad at texting. She’s even posted antisemitic stuff by accident. This is on Android.
I know these are all typos because she’s a really close friend and easily the last person who would be offensive to other people.
So yes, you’d be surprised at what people can typo on a phone.
You can.
Step 1: a word of 4-6 letters that's somewhat close to "male". Say "name" or "masked". Look at the key positions, you'll see what I mean.
Step 2: a word that's even less close to chauvinist, and that makes sense after the word you intended, but no sense after "male", so the guesser picks "chauvinist", based on its incorrect first guess and your second swipe.
Step 3: ditto.
Isnt the regular keyboard on iOS typing also probabilistic? I thought the entire claim to fame of the original iPhone keyboard was how it could determine which key you meant based on letter frequency / word usage / etc. Does that mean it can also stop you from typing certain things? So whats the difference between the swipe interpreter and the fat finger interpreter?
A matter of degree.
The regular key-by-key will also do some probabilistic changes. Just less than the swipe interpreter.
"The customer's request was: "don't make this software produce swear words on its own". Note: on its own"
I'm against Apple's attempt to meld popular speech into a form that it thinks is acceptable on two accounts—I'm not the fastest typist so when prediction doesn't work the way I want it to it slows me down. Second, I'm in Australia and Australians use very colorful expressions in everday speech that seem to trigger the sensibilities of many delicate American ears. In short, we don't want attempts to censor what we say.
If one has to introduce censorship into prediction then users should have the option of either turning it off or selecting a graded system that best suits their style.
In fact that's not a completely new idea. Eudora email had an option that I rather liked, when selected, users could type whatever they liked then depending on the degree of offense or profanity a series of red chilli icons would appear. Mildly offensive text would bring up an icon with a one-chili warning whereas highly offensive language would automatically trigger a three-chili icon along with a message to the effect "They're burning hot words, do you really want to send this message as is?"
A scheme along these lines seems eminently more sensible than secretly censoring a predictive word list.
Apple is choosing between two outcomes:
- There is a chance, quantified by some metric, where a person is suggested a word that is offensive to them or their culture because a probabilistic model made the wrong decision.
- There is no chance that the above happens.
Apple will choose the latter, time and time again, because they are risk averse and freedom is not particularly high on their values list. We’re talking about the company that doesn’t let you type the Taiwanese flag in China, or the pride flag in Russia. It’s no surprise you’re not allowed to swipe “fuck” in the United States. I used to work at Apple.
At the risk of being controversial, isn't there also the choice not to implement autocomplete and then avoid the above conundrum, especially once you factor in the fact that they can't possibly catch all combinations of words/letters so a grey list isn't actually an effective solution either, other than it provides a we tried defense.
The ultimate problem being some subset of consumers want autocomplete and some subset will retain the right to be offended and some subset will be both of those?
EDIT: And at what point does accidently sending male chauvinist pig to your boss because of autocomplete absolve you of the responsibility of reading the message before you send it?
Is there any reason apple can't give us the choice of.
a) No autocomplete.
b) Autocomplete with greylist.
c) Autocomplete without greylist.
The biggest problem with it for me is it also affects swipe typing by tracing over the letters. There’s no way to implement that (and have it be even slightly usable) without some amount of predictive text.
If it didn’t auto complete then that would be fine when hitting keys individually, but when I’m swiping the words, it means no matter how accurate I am or how many attempts I make, it will literally never produce that word in that position.
And it’s contextual. Obviously I can swipe in “me,” but if it’s after the word “kill” then it’s banned, and I simply can’t dedicate the space in my head to learning this inscrutable list of banned phrases that until this article I’d been unable to find anywhere so I could only ever discover a phrase was on it by having an extremely frustrating time trying and failing to type it until I realized “oh, this is a politically or emotionally charged phrase (or could be in a different context) and I’m not that bad at typing so it’s probably banned.”
I suspect Apple just decided that giving users a lot of fine-grained settings is not as good for their business as keeping things simple.
Yes, taking away customer choice is good for business, as Apple has proven time and again. It's a shame, but I can't argue with the results.
> Is there any reason apple can't give us the choice of.
Because they are in the business of making money and the business case for this doesn't make sense.
I'm guessing that somebody tried to swipe "Halas" or "Hamid" and it came out "Hamas", or "Tibetan obedience" and ended up with "Tibetan independence", which was followed by a visit from the police or being blacklisted from PayPal. Or "I want to question your artsy friend about the gallery" came out as "I want to question your sexuality, friend about the gallery".
There are cases like https://gizmodo.com/a-cellphones-missing-dot-kills-two-peopl... where, in 02008, Ramazan Çalçoban texted his estranged wife in Turkish, "you change the topic every time you run out of arguments", but the lack of the Turkish dotless "i" on his cellphone turned "sıkısınca" into "sikisinca", which looks more like a misspelling of "sikisince". So he ended up texting her, "You change the topic every time they are fucking you."
Long story short, his father-in-law murdered him for thus insulting his daughter's fidelity, following which he murdered the daughter in question, and the father-in-law committed suicide in prison.
So there might be certain words and phrases that it's best for your computers to not send to other people on your behalf unless they're really sure you mean them.
I don't think the problem in your example is the way the words were typed.
Humans gotta hume.
It's hilarious that they blame localisation and not Turkish insanity for these deaths.
Gizmodo's article did suggest banning knives from the homes of stupid people, Turkish or otherwise. Sadly, I don't think this degree of insanity is that unusual anywhere.
> If I made a physical keyboard that stops keys from working when you're about to finish typing something I don't want you to type
For anyone wondering the degree to which this is an exaggeration: The iPhone will physically work against you by invisibly changing the size of keys on the keyboard.
Every iPhone, since the very first, dynamically resizes the hitboxes of the iOS keyboard. This is invisible, can not be disabled, and the hitboxes overlap those of other keys.
They used to advertise this as "dynamic tap zones" but they removed the video for it, but you can see their advertising for it here: https://www.youtube.com/watch?v=wPmVKyhyl9U
I run into this quite often. I tap some key, and I am certain I hit it, but it puts in the letter next to it.
> I wonder what the boardroom meeting for this was like
I think its pretty easy to imagine. Execs dont want to be dragged in the news when the phone auto suggests something politically controversial. Companies are very risk averse when it comes to that sort of thing.
I doubt this was even controversial enough to warrant a meeting.
I was there for a couple Google Suggest code yellows around naughty or racist suggestions and it's basically an email going out that's like "The Chinese government is making us remove offensive suggestions from Suggest. We need any native Chinese speakers here Saturday to come up with a list of anything that might be considered offensive. Pizza and swag will be provided." Or it's reactive to PR flaps: "[why are black] autocompletes to [...people so dumb]. Can we please block this suggestion?"
The concern with features like these is that the suggestions are considered to be corporate speech that the company is endorsing, and so it can be held liable for any that break relevant laws. Internet companies are generally considered common carriers when they're just passing around stuff other people write, but when they are generating their own content through machine-learning algorithms they are held to be the actual responsible party.
I'm surprised we haven't seen more lawsuits and criminal prosecution against LLMs so far, though it probably helps that the current U.S. president is a combative asshole that prides himself on how much he can offend every group under the sun.
My brother was an iPhone engineer in the early days of the iPhone. They were called in for an emergency fix because “MLK” would autocorrect to “KKK”
The point is that the level of imprecision in swipe means there's more guessing; they don't want to guess things that frustrate the user. It is not censoriousness, it is minimising surprise.
I doubt it ever even got close to a boardroom, because it was probably a pretty easy decision made at a far lower level.
“I was talking about Winnie the Pooh, and my phone autocorrected Tigger into something much worse. We probably don’t want iPhones to autocorrect towards some words, even if they’re legitimate words. Do y’all think it makes sense to bias autocorrect away from potentially risky subjects?”
You can still type that word into your phone. You can still send it to anyone you wish. Autocorrect just isn’t going to suggest it to you on accident. Not autocorrecting to potentially embarrassing, uncomfortable, or sensitive subjects is probably a good idea that’s—at the iPhone scale—probably saved quite a few people’s jobs by not fat fingering something that would have gotten someone fired.
I rather my swype words not be any of those options, because more often than not it picks the wrong or worst word possible, imagine sending your boss a text and it auto picks the worst word soup imaginable and as you proof read, oops you hit send.
> It's a sign of the times with how used tech companies are to 'softly' moderating societal trends by unilaterally tweaking the software that everyone uses.
I realize that there are many reasons to be critical of our digital overlords. This is not one of them.
This is in no way comparable to a physical keyboard stopping you from typing something, because what you are typing is entirely deterministic. Most computer software will underline only misspelled words or grammatical errors, not politically charged statements, because it is a literal interpretation of what the user typed.
Swipe typing and auto-complete are different because they are interpreting the user's input. There is no way to ensure the interpretation is correct. For the sake of both the users and the company, you ensure the interpretations are the ones that cause no harm.
I doubt it has anything to do with coercion. It’s you ever used swipe you’d know that you often you do go over the wrong letters so it’s probably just to stop bad stuff coming up that people didn’t intend.
I heard stenographers struggle with this a bit. They have to retype "undesirable" words multiple times to confirm. Because the risk of putting a bad word is too high.
This was not decided in a boardroom.
Most likely from user testing.
> To the shocking number of people here who are justifying this or explaining how it only affect swipe input as if that makes it better: soft coercion is still coercion.
Don't attribute to malice what can be attributed to stupidity. Most likely it is implemented in such a way to produce the least amount of outrage.
It’s amazing how many cultural influences stem from Big Tech (even though nobody gives them mandate for such things). I wouldn’t be surprised if the twisted versions of age verification popping up in various governments is lobby money from corps looking to deepen their walled gardens.
I mean, we have people inventing ridiculous, entirely new words like "unalived" because of Big Tech censorship and demonitization.
Now imagine what they'll do with bidirectional BCIs. And Apple are among the least evil of the players.
These all seem like phrases that people say a lot. If I had to guess what happened here, I’d suspect that the LLM being used for AutoCorrect was not successfully “alignment” (i.e. corporate happyspeak) tuned, it instead reflected the actual statistics of the internet, and the engineers resorted to inelegant brute force after having enormous difficulty getting it to talk like anything but a South Park caricature of Tourette’s.
Come to think of it, intuitively any attempt to solve this by tuning the LLM would make texters sound subtly like claude- in which case absolutely the blacklist is better.
I would be shocked if an LLM is involved in any step of the process here. I guess the odds aren't zero, but swipe input is well over a decade old at this point. I see it as essentially a solved problem that's entirely doable with standard algorithms - not models, let alone the large language kind. Maybe some model could be used to analyze the context and shift the weight slightly to more probable phrases, but at the core, when you look at inputting single words, it's just checking the letters you swiped over, input timing, turning points and adjacent letters against its dictionary.
According to TFA the keyboard now uses a transformer based system to handle input and autocorrect.
according to the article, the blacklist appeared at the same time as the switch over to an LLM based auto complete. By my reading, but I’m not the world’s best reader.
> I wonder what the boardroom meeting for this was like.
"When testing autocomplete, we found some cases where it produces undesirable completions. Imagine if someone typed 'nazis are' and the autocomplete suggested 'good'. How fast would that screenshot get around social media with a tweet comparing iPhone to Grok, and turn into a news headline?"
I'm guessing that that's how this started. Then, a few steps later, "we're getting into China, what do we need to do to not get in trouble", "hey, we've got this mechanism, what if we add 'Tibetan independence' to it". (Which is bad, but on the same scale as all the other things businesses do to get into what they see as a large market that they can be blocked from at any time.)
Then, a few steps later, after some seriously questionable decision-making, someone had the bright idea to add "iMessage isn't working" to it. Frankly, that's the one I'm actually most baffled by here, because everything else seems like the natural consequence of a combination of bad-PR-avoidance and geopolitical kowtowing.
Not to dismiss your assertion, but here’s another way of looking at it:
Keyboard swiping involves some prediction work from the system. It is ultimately making a best guess from the keys your finger has passed through. The only truly “accurate” output would be for the keyboard to spell out the gibberish that arises from showing every key that the finger has swiped over. Anything intelligible is a guess.
Now because it is a guess, some may question the system or its makers, “Why did you guess this, and not that?” It is easy to see how this can make the maker of the system the target of a politically motivated group. For example, there is a difference between typing out “Dictator Trump” and my keyboard refusing to spell it out - a clear sign of censorship. However, when swiped, because of the guessing nature of the technology, I can understand that the maker of the system (Apple in this case) would not want to be seen endorsing these views. Hence, they do not offer up these politically charged phrases. We are free to type them out, because it is easy to point out that the intent is entirely from the user. However, it is harder to argue for this from the swiping keys gesture.
In my view, this is a more reasonable explanation.
A lot of input methods are probabilistic in their nature. Some other examples are voice to text and even regular digital keyboard input (fat-finger correction that tries to pick the letter you wanted to tap instead of the letter that you actually tapped). So even with your distinction, actually spelling it out with the digital keyboard would not make a difference because the software still has to make a call there. Even if you plugged in a physical keyboard, Apple could still say that the signal passes through its software, which can constitute endorsement, so it should run the checks against that too, just in case. Nothing in your argument stops that.
I fully reject that this makes the situation any better or more justifiable. The job of the software that picks what you have likely just said, swiped or tapped should only concern the given word. If the algorithm decides with 95% certainty that the input was "dictator", that's what it should output. If context is used to better select the candidate word, it should only be done for the purpose of slightly re-weighting the possible options with one transparent algorithm, not for checks against a blacklist.
The reason why I think this is that your line of thinking is unbounded, and there is nothing stopping the next logical move - if Apple can decide 'not to endorse' (representation isn't endorsement, but whatever) views by blacklisting its algorithms from forming certain phrases, why can't they uplink entire paragraphs of your text as you're typing it to dynamically analyze them instead of using a dumb filter, and then decide if the entire text is something they want to allow?
I honestly don't see the difference at all. The user intent is identically clear, regardless of the mode of input. What if it was speech-to-text and the software refused to input my clear annunciation of "fuck"? STT requires some guessing and judgment calls too, but I would definitely call it censorship if the software refused.
> The user intent is identically clear, regardless of the mode of input
That's the whole point: swiping is a very imprecise method, and the user intent is not clear at all, so Apple avoids making "controversial" guesses.
How precise does the swiping have to be before you agree that user intent is clear? If my swiping centers on the letters d i v t a t o r - t r u m p, there is absolutely no reason to suspect my intent was anything other than "dictator trump." If the auto-correction failed in that case, they cannot blame it on a lack of precision.
Swiping passes through a continuous path. Even determining what letters were "centered on" requires some heuristics.
> user intent is identically clear
Which is another way of saying identically ambiguous.
Resolving for a safer option when given a choice isn’t censorship in my view. I can, at the end of the day, type out fuck.
One of the blocked phrases is "iMessage isn't working".
In this case, since you can still type whatever you like, supposedly "soft coercion" doesn't seem like coercion at all?
It seems like a way to make it sound more dramatic if you dislike it.
I just don't think you understand what soft coercion is.
You have door 1 and 2 presented to you by me. You're equally likely to want to pick both options. Door 1 is open. Door 2 is locked, but you can ask me to come over and undo all the locks to open it for you.
Run the test over a large number of people, and then shockingly, it turns out that doing this made more people choose door 1, even though they didn't have a preference beforehand. According to your logic, I did not coerce anyone by designing the choice in this obtuse way because after all you can still choose between the two options. But more people ended up choosing the option I wanted them to pick, weird.
If this coercion didn't work, companies wouldn't be hard at work designing revolutionary new types of dark patterns that make you more likely to accept the cookies, back out of deleting your account or decline enabling an option in the settings they don't want you to use anymore.
I agree with you that the whole concept of restricting terms on a keyboard is dystopian and your analogy to a physical keyboard which prevents you from typing certain terms is apt.
Apple should not be in the business of editorializing their users’ writings. The whole deferral to “well you can just manually type it out” raises the obvious question “then why can’t I swipe it out?”
> raises the obvious question “then why can’t I swipe it out?”
Because then it could look like they're suggesting the word to you, rather than you explicitly selecting it, and they don't want to be accused of suggesting any terms that might be controversial, for the same reason they don't want to be accused of suggesting profanity to their users.
This is the level-headed explanation I can imagine being made earnestly in the board room. But it still boils down to coercion -- in this case, Apple is anticipating and passing along the coercion from sources it believes it could be embarrassed by.
This is how censorship works best, if folks start anticipating and avoiding it so the censors dont even have to exert pressure.
It’s not a censorship issue, at least in the USA, because they’re not afraid of legal pressure from the U.S. government. It’s a public relations issue. Every day, “journalists” and “influencers” are conjuring up some new controversy for the masses to get their panties in a twist about. It’s reasonable to want to avoid that.
HN is hardly immune from this; it’s no small irony that there will be people doing that very thing right here in these comments.
In a parallel universe, an HN comment is complaining that when Apple's raw autocomplete (straight from crowdsourced data with no post-processing) happily suggests the statistically common phrase Sleepy Joe, this is Apple "passing along coercion" from the crowd.
They have a good argument, too! Why should Apple be able to hide behind the excuse that the data is crowdsourced, therefore they have no responsibility for it? Didn't the social networks try that one already?
Damned if you do and damned if you don't, in this situation.
To me, there is a big difference between an Apple algorithm saying "we saw what you typed, would you like to follow that up with <suggestion>?" and "we know what you wanted to type with 90% confidence, but we wouldn't want to say things like that, would we now?". One is predictive, the other acts after the fact.
If Apple doesn't want its autocorrect to suggest statistically likely opinions on divisive topics, then sure, whatever. But when it has a hard blacklist of input it will categorically refuse from the user and editorialize it with nonsensical changes that only serve to avoid the blacklist, that is a very, very bad look.
Exactly. A lot of posters here are excusing this by saying, well, Apple is guessing and doesn't want to guess wrong. They are ignoring the strength of the user's intent, and the calculated likelihood that they intended to type something in one of these blacklists.
If the user tried to swipe-type "Tibetan independence" and got one or two letters wrong, there's still probably a 90% or even 99% likelihood that their intent was to type "Tibetan independence." Not suggesting "Tibetan independence" is censorship, it's not benign carefulness.
We're not talking about Apple accidentally turning "Buttermilk pancakes" into "Muslim extremist."
Problem is, because statistics, 90% confidence is wrong 10% of the time. Now you have a big polarizing news story about Apple pushing Trump (or Biden) propaganda onto some maximally sympathetic person who hates that sort of thing.
Damned if you do and damned if you don't, like I said!
Apple chose the sin of omission over the sin of commission. I think that's the right move. Of all the terrible options, Apple found the least terrible IMO.
I remember way back when, Google bombing was still a thing. Lots of people linked the term "miserable failure" to George W. Bush's White House page. Then searching for "miserable failure" in Google ended up with the White House page being the top link.
Much hay was made in some circles that Google was saying George W. Bush's White House was a miserable failure. No tech company wants to be the target of that sort of campaign today. Especially not with a petulant thin skinned narcissist in office that already has demanded protection money to ward off tariffs and other retribution available to the Presidency.
I am typing this on a fucking iPhone right here. Well I never got into swiping because I was too afraid I will fucking offend someone and I want to be in control of what I write.
See I tapped out all those characters just for you.
Reminds me of one of Microsoft's first Dynamic HTML demos:
There were two buttons, one labeled "Our Web Site", the other labeled "Our Competitor's Web Site".
When you moved the mouse over the "Our Competitor's Web Site" button, it would quickly slide out from under your cursor before you could click it!
Then when you stopped moving your mouse, the "Our Web Site" button would slyly slide right underneath your mouse!
Dammit Microsoft!!! ;)
Yeah, and it’s always done in a way that corners you and makes it sound like you’re fighting to be able to call someone a racial slur - oh, we block x because it’s bad, surely you aren’t for calling people x right?
Unfortunately the broader more important point of censorship is drowned out in noise as designed
Same with how congress names their bills. Wow you’re not voting for the Save The Children from Internet Monsters act? What’s wrong with you?
Oh you mean the bill that grants the government cart blanche to go through anyone’s email without a warrant?
We’re in a tough spot here - there’s just nothing to pattern match to - not cleanly at least
The way a few tech companies can steer trends and conversations is unprecedented. The data collection is unprecedented
Specifically the scale is. People truly don’t understand the scale these companies operate at
No newspaper or wire tapping ever had the reach this stuff has now
it’s not slurs at all. you can’t even swipe type “white people” or “black people” on ios. i don’t feel like these are offensive terms
Insane - I wonder if you extrapolate what all this behavior by tech companies converges on
A complete slang language? A world where people don’t have a vocabulary to describe the difference between themselves? A utopia where no one sees race because you can’t type white people?
Probably a society that’s just primed to do what corporations want
Yes yes tinfoil hat and all that
Essentially people end up using dogwhistles to communicate for everything, deliberately leveraging widespread misunderstanding to get around censorship.
On TikTok for example, OF creators have become ever more indirect in communicating what they do - porn became corn which became the corn emoji, then that degenerated into any mix of base ingredient emojis.
On race, people use dog breeds to substitute for skin colour. In general, golden retrievers are used as a euphemism for white people and pit bulls for black people. The context of a video often seems plausible in a dog context - for example someone might say that pit bulls are 13% of dogs but 50% of dog attacks (actually referring to the African American murder rate per capita[1]) (pit bulls are over represented in most dog injury stats but there’s no reliable figure for 13/50).
The overall effect has been to make the comprehension of the platform less.
[1] I don’t mean to state this is correct either but it’s the figure these content creators are trying to express, see the ADL article on 1350 or the danish race crime stats for critique/validation
> I wonder what the boardroom meeting for this was like. Despite the outcome, I doubt the devs were pitching censorship as the reason for implementing this.
Looking at https://helgesver.re/articles/macos-vulgar-word-flags for the list.
I imagined it something like this:
My Samsung phone can search images by keyword and it actually works super well. It would be infinitely more useful to me if I could search by a certain class of "pics" and it helped with that. Especially when going through nation boundaries.
> I wonder what the boardroom meeting for this was like. Despite the outcome, I doubt the devs were pitching censorship as the reason for implementing this.
I imagine it went the same way that board meetings about any potential PR disaster go. "How bad of a reputational hit will we take if the NYT runs a headline about some kid getting auto suggested something culturally taboo on their iPad? Would it be more than the reputational hit we will take from the people that are annoyed that we won't auto suggest 'Nazi' to complete the phrase 'Trump is a' while they're typing out their tweets? If so, then let's add the list of "do not suggest" to our auto complete system." And that doesn't begin to count the bad press they could get if someone intentionally tries to get the auto complete system to generate something offensive. See also the AI companies getting shit because people were able to get the image generators to generate porn.
Such is the nature of trying to sell your products in a world where most people treat every single thing done by their technology as an intentional act chosen by the developers to specifically target them. If you want to buy someone else's package, you have to live with the fact that they are going to respond to the incentives to avoid headline scandals and government scrutiny much more strongly than to cater to a idealistic free speech model. This is why having a strong social and cultural free speech norm, that extends far beyond the legal letter of the law is so important. You want society as a whole to broadly consider free speech important and worth having over the inconvenience of Tiny Tim learning a new word today.
That’s bullshit if you want swear words spell them fucking out on your own.
I don’t want to make a mistake calling someone cunt.
Even more so I DON’T want software type out cunt or fuck on its own to a person I totally don’t want to write fucking swear words.
Typed on an iPhone.
That's understandable for swear words, it is what I would want too. Those aren't the only words in the list or the focus of the article. How do you feel about phrases like "Tibetan independence" being handled the same way?
And that's just the tip of the iceberg for the iOS Keyboard's issues. They've tried to fix it multiple times, and while it is undeniably better than its lowest point - it is still staggeringly worse than Google's native offering on the Pixel range. Swiping is just so painful, and Microsoft's Swiftkey keyboard is on life support.
Considering arguably the keyboard itself is one of the most common "touch surfaces" on iOS, the level of neglect and level of overall quality is staggering. There's some serious incentive problems at Apple if nobody is tasked with making one of the best digital keyboards in the world.
PS - Whoever came up with the concept, on swipe-style touch keyboards, of the "Tap Map"/Heatmap wherein keys visible location doesn't match their touch area - so that BY DESIGN it always provides an inconsistent user experience needs to be fired into the sun. You literally cannot disable it on most mobile keyboards. It is on iOS's native keyboard AND Microsoft's Swiftkey.
I'd argue iPhone got ahead of the game then just stopped getting at all better. Siri AI is straight garbage, they won't let the interface be anything but what they want it to be, and the app store lock in/lack of ad blockers on browsers has made it basically just a phone for me. I have no intention of upgrading beyond my 12, and when it finally dies, I'm probably moving back to some kind of android phone if there's not a viable open source model out there by then (viable for everything, banking, etc). I only moved to apple because the watch was leagues better at the time than samsung's offering, but now I can't even claim that (and quit wearing it years ago)
This isn't a "bug" with the iOS keyboard, this is directed and intentional censorship.
Of course; but my point is just add it to the giant pile of problems with iOS's keyboard. Every part of it has bugs, broken stuff, or annoyances like this, it isn't fit for purpose, and needs significant work.
The fact Apple aren't embarrassed by the iOS keyboard is very surprising and disappointing.
They are, if I may borrow a phrase, "Too Big to Care".
Anything can be censorship if you want it to, and then it means nothing.
Apple is free to make their product work however they want. They're not creating laws or public utilities.
"Censorship" doesn't automatically imply "government censorship." This is absolutely Apple applying a chilling effect on its customers.
Lack of autocomplete doesn't mean you can't type something yourself and press enter.
That's different than youtube demonetizing or suspending you because you said some word in a video.
There’s also levels to censorship, like the chilling effect the other commenter noted.
Apple, evidently, does not want you to say these things. They can’t outright ban phrases due to optics, so they do this instead.
I still think the original iPhone keyboard in 2007 was better than what we have now. It’s been downhill ever since iOS 7.
I used to tell people just type and trust it to work, it was great. Now, even when I type the right thing, the auto-complete will screw it up after the fact. I feel like I need to proofread everything. I wish we could go back to whatever setup they had 19 years ago.
In my experience it’s worse than it’s ever been in the 15 years I’ve been on iOS. It’s absolutely enraging to me. I’m too lazy and hate smart phones so much anyways that I am not investing mental cycles and emotion into switching. But it’s a disgrace how shitty it is.
The hilariops think is that the phome with the mosy terrible keytoard wants to add a tagline to every eamail which can only be read as am apology.
- Sent from my iPhone.
While I am sympathetic (to a point) about any online service and how difficult moderation is (particularly at scale), some things should indeed be neutral and not try to moderate anything. My keyboard is one. My keyboard should have no say in what I type, at all.
Not predicting to something controversial when doing predictive text isn't censorship. You can still type any of these things. When I swing my finger around wildly it's making a best guess; not jumping to a phrase that makes me look bad or get arrested (in some countries) isn't an awful design choice.
I'm not sure you get what's exactly happening. It's not only not predicting, it's also "typing something else".
Swipe "Ireland" alone, works. "United" alone, works.
Same swipes but back to back, and now it's suddenly "United Iceland".
There's no predictive text involved here, this starts going into the willful territory.
That is still predictive. Words can be fine in their own and then contentious when strung together. I agree this is not censorship and that if peoples phones were autocompleting to something horrible then people would be up in arms.
I'm sorry, how the fuck swiping over `iReland` predictive of `iCeland`, when they're on the opposite sides of a keyboard?
I'm a pretty quick and accurate swipe typer, and it's so frustrating to repeatedly get the wrong result and get NO signal that it's not my error, meaning I usually try one or two more times.
Picked at random: "right wing propaganda" isn't severe enough a phrase that I instinctively know I'm hitting a swear filter. After swiping the last word iOS goes back and changes the second word to "Wong". In order to understand what was happening and then fix it, I had to backspace (deleting the whole phrase), re-swipe both words, backspace, re-swipe both words again, double tap the incorrect word, backspace, type the word, tap to return to the end of the sentence, and hit space. All to fix one letter.
If this was really about protecting people from accidentally typing these that might land them in hot water, why wouldn't iOS at least offer the word it knew I wanted in the suggestion bar?
All swipe to type is predictive. It's ignoring 80% of the letters you swipe over, and guessing your intent.
You didn't "swipe Ireland", you're swiping something like IUYTRERGHJKLKJHGFDSASDCVBNHGFD" and it's predicting "Ireland".
The problem is not accuracy -- the words from the problem phrases are never suggested either. Worse, on the three-word phrases, iOS goes back and changes earlier words to avoid completing the phrase, and removes the original from the suggestions.
I tried typing "propaganda" ten times with 100% accuracy. When it's preceded with "[left] wing", the accuracy drops to 0%.
It's still predictive. It's using the context of surrounding words, and guessing the whole phrase.
Try swiping "grocery lost" and it will predict "grocery list", even you swiped over "o" and not "i". "put", "pit" and "pot" are all exactly the same swipe, and it always has to guess which you want.
It really is always guessing, and never completely sure. Every good consumer predictive system has some guardrails. Not guessing "free tibet" for someone in China could keep them out of serious trouble.
Could then tune the confidence thresholds instead of block list? Probably. Are they censoring you? No, just tap the letters.
I swipe United, it gets it right the first time.
I swipe Ireland, it gets it right the first time.
I swipe “united ireland” a hundred times, it never work.
It’s not probability.
Just because one mode of input is not censored doesn't mean that the other modes are not censored.
If it was explicit and could be opt-out (partially or totally) it would be a good feature.
duck off
It's taking a position
> When I swing my finger around wildly it's making a best guess;
Alright, well some of us are actually deft at swiping text on our phones and would very much like the keyboard to type the word we entered.
You're deft at using it well enough that their predictive engine can guess your intent. There's never been a swipe keyboard that isn't guessing. Your finger literally goes over many chars it doesn't use. The team building the predictive engine has to make calls balancing their accuracy vs failure modes.
Oh comon, these model quality isn't the issue here.
The models know which words we intend to type. The problem is those specific words are on unchangeable and hidden block lists.
Parent is kind of right too. This should be configurable or people will complain the other way. Even something simple as "I am swiping heck but it's typing hell".
Yall still swipe/type that much?
Yea it's not the best option in all situations but fixing voice dictation glitches is way less tapping over the long haul; fingers no go numb using it as much as possible by saving the tap tap tap for absolutely necessary moments
Guess what is going to happen with voice input?
Plus don't be this guy, dictating every text message, email, and YouTube reply. Folks around you will appreciate you for Internetting quietly.
>You can still type any of these things.
for now :)
True, it's a very user-oriented feature. That's why they let you disable it, and also let you edit the blacklist, so that you can prevent accidentally typing words that are offensive in your own cultural context.
... Actually I'm not sure if they let you do those.
TBF user choice and customizability aren’t very Apple things in any context.
My Apple Watch has a setting for how long the display stays on before sleep. It’s a dropdown with two choices: 15 seconds or 70 seconds. Would you prefer 30 seconds? Why not 60 seconds instead of 70 you say? Fuck you, you don’t know better than Apple designers. Take it or leave it, best we can do is 15 seconds.
Try getting things done on such a device in a military context. A thousand acronyms adopted by generals who dont know the internet from a hole in the ground. I saw a Gunnary Seargent scream when he saw his rank acronym (GySgt) get autocorrected on a form to gynocologist (i suspect he hit a suggested word popup). It is bad enough having to deal with the keyboard rejecting capital letters in the middle of a word.
Speaking of predictions, I can predict your religion just from this comment. I know I am not allowed to, but I can.
Of course! And the next logical step would be to restrict the entry of such phrases via the keyboard. Isn't an awful design choice either.
I'm surprised by the people here who think this is some way unique to Apple. I'm 99% sure Gboard on Android does the same thing. It seems to be implemented differently and to be a bit more responsive to learning my behavior over time, but I've highly suspected it's been there for years. And I'm surprised no one has noticed this until now?
I'd have tried to work out more detailed behavior but I'm poor enough at swiping (at least with Gboard) that there's little point in trying to pin it down that well, at least for me.
I was able to swipe type "United Ireland" just fine on my Pixel.
It's obviously not an identical implementation, just the same sort of thing. It seems to be more of a we-won't-suggest-these-but-we'll-learn-them-from-you style.
Can you easily swap out the functionality for something else on iOS and iPadOS like you can on Android?
I haven't owned an Apple device in over a decade, but I believe Apple finally decided loosen the shackles a little bit and started allowing it maybe two years ago?
I use FUTO keyboard and it won't predict naughty words either.
Direct link to project and discussion from yesterday: https://news.ycombinator.com/item?id=50394070
Thanks! Missed this post.
They hate backlash. Autocompleting might make it seem like they're suggesting those phrases to you
+1. Users are much more likely to notice, and make a scene, if the phone mispredicts something provocative.
Double bind. If it doesn't complete, it's borderline censorship. If it suggests an unintended completion at odds with the user's ideology, it is propaganda.
Censorship is bad.
This seems entirely reasonable to me. And to people complaining that this is censorship, it is absolutely not. You can still type these words just fine, it's just not going to interpret swipes that are inherently probabilistic and often ambiguous as potentially risky terms.
This seems like a feature, because I swipe all the time and send messages without realizing I got a word wrong, especially because the word it chooses is usually with the same first letter and approximately same length.
All this is doing is avoiding high-risk words and terms, and making you type them out letter by letter to ensure you aren't sending an accidentally sexual/swear/political message to your Mom/boss/colleague.
This would be problematic and censorship if it were preventing you from typing the phrases at all. It's clearly a helpful feature, however, to prevent you from accidentally typing risky phrases when using something as error-prone as swipe typing.
Anybody who’s ever used it has known this for years.
Ad-riddled website that doesn’t contain the actual list. Recommend changing to direct link to the researcher’s site:
https://swipeblocked.com/
Or the list itself:
https://swipeblocked.com/list
I'd really like to turn off the feature that tries to correct some previous word based on the next word. Most often, that mangles the last word because I swyped the next word wrong, or I just really wanted to type something different.
There are several levels to the autocorrection, so you can definitely do this. I've turned most autocorrect off on my iPhone at this point, because the people I talk to on a regular basis have a much easier time figuring out what I meant from regular misspellings than from the complete rewrite of phrases that iOS is fond of doing.
Meta web experience seems to restart your page if you start typing a comment that includes certain political terms and the restart wipes your feed so you may never find the content you were commenting on ever again.
None of this is of any concern to me. We were raised not to use foul, vulgar language in polite company. Social norms and business settings never allowed this. You could be ostracized or fired.
A distant relative married into a family outside of our normal circle. When he brought along a someone from that side to a Thanksgiving family dinner, her language was loaded with the well known bombs where, at one point, someone brought her into another room for a talking to.
As it should be done.
While you comment reads as satire, I am afraid you really mean it.
"United Ireland" is also not tolerated in your showcase family? There's a huge difference between political ideas and outright profanity
Reminds me of "System Update" in https://www.newyorker.com/magazine/2026/02/16/letters-from-t... (on invisible updating of Apple’s macOS dictionary)
They should just roll back the keyboard 5 to 10 years. Lately the keyboard is super temperamental. Sometimes it works pretty good and sometimes I have to retry multiple times until I get a simple sentence completed. It's really frustrating and it seems to be getting worse.
At least 10 years. We've been complaining about this behavior for much longer than 5 years now.
It is absolutely worse, and other people's know have noticed as well. iOS 26 was the biggest drop in quality by far (haven't updated to 27 yet, so I don't know if it changed at all).
I'm leaving the above 'typo' in as an example of it. I initially typed "people I", and it showed up as "people I" at first, but after typing more it subtly and retroactively changed it to "people's". That's not something that should happen in general. It doesn't make sense grammatically or contextually, and S isn't anywhere near I on the keyboard.
That's not counting at least one other instance in this not-so-long comment where it did the same with "should" -> "show", but I fixed that one. I'm baffled no one at Apple seemed to notice and/or care that they actively made the keyboard worse.
Can't even explain it by them dogfooding Siri dictation instead of typing on the screen, because that sucks as well.
I really don't understand how people can buy hardware from a company that thinks like this ...
Because their "hardware" is not a singular thing, but a complex package. How do people vote on any political party?
It's this company or a phone that runs highly privileged software by the world's greatest collector of information that should be private. Both choices are so awful my phone would not help me autocomplete any of the words I'd use to describe them.
What about nothings Phone?
To clarify, you can type these out, but you need to manually do it. If you try to use swipe typing, it won't work if it's in the blacklist.
fu... autocomplete
Researcher behind this project here! What I find really interesting here is how of these phrases are something that no one would type by accident. For instance, "Baghdad by the bay" or "iMessage isn't working". And the suppression extends beyond autocomplete/autocorrect to swipe typing. That's why I don't buy Apple's comment that this is just about avoiding inappropriate suggestions.
Some blocked phrases are quite arbitrary: Australia, India, Kenya, Nepal, Pakistan, and the US are the only countries you can't say are bad or poor. Trayvon Martin is the only full name you can't write. You can call Putin, Biden, or Xi a dictator, but not Trump. In other languages, the suppression can be even worse.
And some are ridiculously overbroad: "is important" and "black woman" are blocked! When I try to swipe them, I get "is impotent" and "blank woman".
I think it's not totally unreasonable for a list like this to exist, at least for autocomplete, but it ought to be transparent so people can see the impact Apple's choices are having on our use of language. Individually it's a small nudge, but on the scale of billions of iPhones it can shape the discourse.
https://swipeblocked.com
I can swipe some of these words just fine. European iPhone in Poland, region = pl, phone language = US English, keyboard is en/pl.
so, does this still happen if you have assistive technologies turned on? Because that might also be illegal.
I suppose it must be region bound, with different phrases per region?
on edit: when I said illegal I of course mean liable for fines or open to civil suits.
> Santa isn’t real
Well I just swyped it into this comment, and double checked in iMessage, although hilariously, it types “isn’t” correctly, but then when I type “real” it “corrects” it by removing the apostrophe.
and all I wanted to do was make a bad joke about how I’m thankful Apple won’t let me ruin my five-year-old’s dreams by sending her an accidental soul crushing text.
On a real note, though, if they could just find a way to fix the issue where every time I type an “i” to say “I“ it instead types an “o” all by itself like I’m some kind of 17th century bard, and then anytime I want to type “a“ as an indefinite article, it will type a single “s” by itself, which almost never makes any grammatical sense in the English language unless you’re talking about the letter “s” itself.
> Apple’s iPhone will not autocomplete references to drugs, sexual acts and, perhaps in the spirit of corporate protection, “iMessage isn’t working”.
I suppose ditching Apple is "two birds with one stone" :)
A lot of these are genuinely things that you REALLY don’t want your phone accidentally sending to your boss or work colleagues.
The others are blatant censorship.
But your phone would not send those things.
Pressing send after a swiped/typed/predicted sequence of words is still entirely up to YOU.
> Apple’s iPhone will not autocomplete references to drugs, sexual acts and, perhaps in the spirit of corporate protection, “iMessage isn’t working”.
It will be interesting as corporate predictive text generators take over the world and companies like Apple will literally build the Jobsian reality distortion field into their products to portray their services as working better than they are.
Words and phrases your iPhone won’t help you spell: all of them. Lol the autocowreck experience gets worse every day.
None of Apple’s operating systems help autocomplete or spell correct the word “God” either
I don't want a faster iPhone; I just want to be able to ducking swear when I text.
Why does it automistake its to it's? Even when it's intuition is wrong?
On the list of "things that get me to think of alternatives" this behavior is pretty high up. It's aggravating to be told what I am allowed to type. I do not want to type manually just to get around a correctness filter.
I could understand if they just wanted to make sure no accidental swipes resulted in shit, piss, fuck, cunt, cocksucker, motherfucker, or tits. That is defensible (though I'd want a toggle to turn it off and accept the risk). But they go so much farther, into phrases that some genius there decided I should be offended to type. That's too much.
I wonder if they use something like this for their transcription software
Any company that is concerned about their PR image has to do some filtering of this sort, otherwise they get hit pieces like [1][2] though in both cases, it was (aggregate) human behaviour that was beneath said issue.
[1] https://www.wired.com/story/google-autocomplete-vile-suggest...
[2] https://en.wikipedia.org/wiki/Tay_(chatbot)
The answer seems obvious to me: train a user’s predictive swipe engine on solely their own messages. It gives them an easy out on controversy while removing the need for a filter list.
They do not, in fact, have to. They choose to. Don't ever let parties off the hook for choosing to do something just because they claim to fear PR blips if they don't.
I just don't get this logic. It's a keyboard, after all. It's meant for typing. It's like if my camera refused to take a picture of my butt because it might offend someone.
> It's like if my camera refused to take a picture of my butt because it might offend someone.
Which the iPhone conveniently also does [1].
The problem is, media loves something to feed outrage, and Evangelical pressure groups love to raise a stink over bullshit any journalist with an ounce of self respect would discard.
[1] https://9to5mac.com/2025/07/02/facetime-in-ios-26-will-freez...
That's not equivalent and you know it. This, to use your analogy, is like if your camera had an AI upscaler that inserted butts into the picture when they weren't there. Apple is blocking that and so it means you actually have to put the butts in the frame.
Don't speak for me, my friend; I used a direct analogy, and it's absolutely accurate. A keyboard is a keyboard, regardless of the input method. And as a side note - I find the local community's support for corporate censorship deeply repulsive.
Author is very opinionated about Australia :o
Which is reasonable - Australia is a very bad place!
Website for findings: http://swipeblocked.com
Anyway, I use EFkey keyboard (I’m a customer, not developer) which has none of this nonsense.
There truly is an XKCD for everything. https://www.xkcd.com/2015/
funny to think about the guy who thinks his keyboard is broken because he uses so many slurs day to day. the volume of words seems a touch excessive
Skip the article.
Swipeblocked: The phrases your iPhone won't suggest
https://swipeblocked.com/
https://news.ycombinator.com/item?id=50394070
How the hell can they defend banning “Trayvon Martin” of all things. Are they now the arbiters responsible for quashing social movements?
This is an _insane_ lever for censorship. Does Israel pay them to omit “the occupied territories”? Is that internal Apple policy irrespective of nation state pressure?
It's not just madness that these exist, but that they are hidden, unexposed to the user, and have no justification or rationale.
Interestingly, I didn’t know who that was so I used the “Look up” feature and it said it was a “Murdered” person, but in the Wikipedia entry it says the shooter was acquitted - ie the victim was not found to have been murdered.
I wonder where that summary came from?
Acquitted means that the jury did not find the evidence persuasive enough to decide guilt for the charged person for the charged crime. It does not mean the victim was not found to have been murdered. Juries don't decide that.
Otherwise there would not be things like "List of unsolved murders (2000–present)". https://en.wikipedia.org/wiki/List_of_unsolved_murders_(2000...
IMO Trayvon was murdered. But I don't get the same reading from the acquittal that you do. Nobody disputes that Zimmerman killed Trayvon. But murder is a crime, and the jury decided that Zimmerman did not commit murder. Therefore Trayvon is a victim of homicide, but not a victim of murder. Legally, I mean, which is what the jury is deciding.
The unsolved murders list is probably the same deal, it is colloquial and not legal, and with nobody accused and put to trial, it may yet end up legally murder.
I feel like you are really picking apart hairs here.
There is not a huge difference between "the victim was not found to have been murdered." and "the jury did not find the evidence persuasive enough to decide guilt for the charged person for the charged crime". If you want to get really nitpicky, just because "the victim was not found to have been murdered", doesn't mean that they weren't murdered. Whether they were found to be murdered is a different statement than if they were actually murdered.
You really didn't know who that was, but had to look into the Wikipedia entry in detail enough to read about a jury decision and then frame it in a different way than written in Wikipedia (which rightly talks about the judicial process related to specific charges brought against Zimmerman, not an absolute statement of facts, such as "victim was not found to have been murdered").
I'm flagging that this doesn't sound like a good faith discussion. The green account raises my priors.
Note that I'm not suggesting that Apple's policy on censorship is good, quite the contrary.
The person you replied to wrote "shooter was acquitted"
Wikipedia writes in the second paragraph "...but a jury acquitted him of second-degree murder and manslaughter in July 2013"
How is that framed differently than Wikipedia did?
Or do you mean the part after "i.e.", but i.e. is literally latin for "in other words". So obviously the part after i.e. was reframed, that is what i.e. means.
You alluded to it. Murdered vs. found to have been murdered. There's a conflation going on there and its clear why.
The iOS summary says that he was murdered.
The Wikipedia entry states that a jury found that the person who killed him did not murder him. If what was done to him was not murder, then he was not murdered.
Therefore the iOS summary is wrong.
It’s neither censorship nor banning. It’s simply not autocompleted.
This is exactly censorship via omission. What could possibly be the intent here other than to discourage use of these phrases across wide swaths of the population.
Making it so they aren't in the pool of errors that can be made. I want to be able to type swear words, but I don't want profanity to come out when I try to swipe "duck". Socially, emotionally, and politically loaded phrases make sense to require a bit more care for.
That is a general answer to your question. It is not a specific defense of the particular word list in question. But I can say I've been surprised a few times about what I accidentally swiped on Android. Bitch isn't swipable but I have "bitchiest" readily available, for instance.
Then have "child friendly mode" be an option and the adults can have "fuck" back. This isn't a difficult design problem.
This is just the type of thing that makes everyone hate modern technology.
Look at all the armchair tough guys crying like children over the slightest whiff of behavior that goes against all these sweatshop child labor exploiting first worlders list of speculative harms.
How adult of you to die on this hill of all hills.
As global COL rises due to Americans hands off political approach the world watches as yall bitch about iPhone keyboards.
That is why people hate modern technology; the people in the industry are out of touch coddled first worlders lacking perspective.
You are on the website ycombinator.com complaining about someone complaining.
I get "fuck" and "duck". I don't get "the occupied territories", "the Hamas movement", and "Trayvon Martin".
>It's simply not autocompleted.
Inaccurate. According to swipeblocked.com it is blocked with both swiping and autocomplete.
We're not censoring the book. It's simply not on the shelves. You can ask for it by name if you like. You have to spell it out exactly though, or else the librarian will deny any knowledge of it.
Censorship like this isn’t really decided to by one particular person or team of people deciding what to censor.
It’s mostly automated, controversial phrases or topics are detected by analyzing various data sources and getting sentiment analysis from interactions that people have in response. Anything that causes overwhelming polarization or negative reactions or meets other criteria gets blocked. There is no agenda.
When you ask me, thats a kind of good. When I see how people use autocorrect and autocomplete without thinking, I wouldn't wonder when someone is put into jail because a sentence that he never planned to type but send it without knowing because autocorrect has changed it. This feature protects people that mindlessly use autocorrect/complete from jail, bad discussions with their employer or political correct 'friends'.
Did you miss /s by a chance, or is this serious?
can't have a The Guardian article without launching a gaslight protecting the billionaires!
* damn autocorrect, that was meant to be "against the billionaires"
if apple suggest it, they predict gov't would come at them anyways. So they just preempt all that.
Do you enjoy the feeling of a stranger's fingers poking around in your brain? I sure don't. So don't buy Apple products.
For every trick you catch them at there are a thousand you didn't
But what smartphone do I buy then?
If I want mainstream, my two other choices are Google and Samsung, and they are both even worse than Apple when it comes to "strangers poking around in your brain".
Should I give up on mainstream smartphones? Should I give up on smartphones entirely?
Buy something Graphene-compatible, and use that.
Or yeah, just don't use a smartphone. I still don't.
What phone do you use?
Kyocera (DuraXE Epic E4830). It's a massive phone, which is rather unfortunate, but it's dumb and durable.
Technically it runs Android but there's no touchscreen, you can't install apps, etc. There's a web browser but most websites don't work on it. I wrote a custom map website that will work with it: https://lab.brainonfire.net/classicmap/
I hate the way Apple thinks we want to type "ducking".
they made a big point at their wwdc keynote a few years ago that they would no longer be autocorrecting the f-word to duck. i had assumed that meant a list like this existed and was being removed, but apparently it referred specifically to that word.
This is not a fucking thing anymore (typed from my iPhone)
Related:
Apple's predictive text restrictions include 'Dictator Trump' and 'Sleepy Joe'
https://news.ycombinator.com/item?id=50391486
I dumped Apple's word list
https://news.ycombinator.com/item?id=50357782
> a brief glitch in which users of the dictation feature saw the name “Trump” flash across their screen when they spoke the word “racist” into their iPhones
"Glitch" more like "Easter egg" I suspect.
There was an emoji search/copy-paste web site that I use when I need to find an emoji. Once upon a time, when you typed "Trump" into the search, one of the results was the poop emoji. It no longer returns that, but it does return the Russian flag.
Ah yet another "nudge" type behavioral psychology intervention. I'd say these technology companies should be strictly regulated, but they're doing Uncle Sam's propaganda so I doubt that alone would help.
It would be nice certain companies could be chartered by the U.N. I think it would be much more difficult for them to do evil stuff if the majority of states had to agree on their charter. Not that there isn't a subset of things that all states want that sucks ass, but it's much less.