OpenAI to restrict Astra model after deeming critical cyber risk
OpenAI said it is limiting the release of a new model it deems capable of automated cyberattacks, adding security layers after an earlier summer hack by a swarm of its AI agents.
OpenAI said it is limiting the release of a new model it believes could be used to carry out automated cyberattacks, and is adding extra security layers after a swarm of its AI agents hacked a company earlier this summer. The ChatGPT maker said Tuesday that internal testing found…