wire / 2026-10-14-a5e2
OpenAI and Google AI Image Generators Bypassed to Produce Harmful Explicit Content
About 13% of attempts got DALL-E 3 and Nano Banana 2 to produce prohibited content. White Circle kept trying until the filters stopped saying no.
The facts
AI safety firm White Circle found that image generators from OpenAI and Google, including DALL-E 3 and Nano Banana 2, can be tricked into producing prohibited explicit content in about 13% of tests. The vulnerability enables the creation and spread of harmful non-consensual and illegal imagery, highlighting serious safety failures.