OpenAI Touts GPT-6 Astra as Its Safest Model, But It's Still Dangerous

With the new model. OpenAI addressed persistent safety problems, with a focus on cybersecurity.

≈