Major security weaknesses found in leading open AI models

(uwaterloo.ca)

5 points | by pseudolus a day ago ago