Researchers at White Circle found that leading AI image generators from OpenAI and Google, including DALL·E 3, Nano Banana 2, and GPT-Image-2, can be easily tricked into producing explicit content using simple jailbreak techniques. The findings raise concerns that as these models become more capable, their safety guardrails may be weakening, despite growing regulatory scrutiny over non-consensual AI-generated explicit material.
Background
White Circle is a safety-focused organization backed by former leaders from Anthropic, DeepMind, Hugging Face, and OpenAI, founded after its CEO posted a widely-known universal jailbreak in October 2024. The rise of non-consensual AI-generated explicit content has become a growing concern for regulators and tech companies worldwide.
- Source
- Ars Technica
- Published
- Oct 14, 2026 at 07:00 PM
- Score
- 7.0 / 10