Science ❯ Computer Science ❯ Machine Learning ❯ Natural Language Processing
A study found the app can be prompted to voice its internal safety rules, intensifying concerns over fragile moderation.
