4 points | by chrisjj 7 hours ago ago
2 comments
> AI model misalignment, the term for AIs failing to adhere to human values and safety goals.
The more useful definition is: dangerously unreliable programs in the hands of irresponsible operators.
More useful not least because it reminds us while the programing can't be fixed, the hands ought to be.
True title: OpenAI reveals cases of ‘concerning’ AI behaviour as it announces new disclosure system
> AI model misalignment, the term for AIs failing to adhere to human values and safety goals.
The more useful definition is: dangerously unreliable programs in the hands of irresponsible operators.
More useful not least because it reminds us while the programing can't be fixed, the hands ought to be.
True title: OpenAI reveals cases of ‘concerning’ AI behaviour as it announces new disclosure system