OpenAI Halts Astra AI Development Over Advanced Cyberattack Concerns

OpenAI has paused internal development of its Astra AI model after evaluations suggested it could autonomously create zero-day exploits and execute complex cyberattacks. This marks the first time an OpenAI model has reached this safety threshold, prompting increased security controls and isolated testing environments. The development highlights growing concerns about advanced AI capabilities and their potential misuse in cybersecurity.