OpenAI says it's "pacing model development" as AI cybersecurity risks grow too dangerous

OpenAI says it's "pacing model development," partly because the upcoming "Astra" model may be close to gaining critical cyberattack capabilities. The company paused reinforcement learning for two weeks, its "largest planned frontier RL run" remains on hold, and workloads that haven't met new security requirements are suspended. The Hugging Face security incident and "rapid progress in our internal research" also prompted the slowdown.

Since then, OpenAI says research environments have been hardened with better network isolation and stricter sandboxes. A new monitoring system alerts within 30 minutes of detecting suspicious behavior, using roughly 20 percent of supervised inference compute depending on workload.

The company plans to expand its Preparedness Framework and invest more in alignment research. It has, however, disbanded the team behind that framework, shifting responsibilities to other teams.

Critics will likely keep accusing OpenAI of fear-mongering to buy time and attention. The independent government agency AISI has documented similar harmful model behavior, lending some weight to OpenAI's claims.

AI News Without the Hype – Curated by Humans

Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.