18 August 2026

OpenAI halts some AI development citing cybersecurity risks

• OpenAI paused two major research efforts, including work on an upcoming model called Astra, citing concerns the model may gain dangerous cyberattack capabilities.
• The company suspended a two-week reinforcement learning run, a technique for training AI systems to improve through trial and error, and halted workloads failing new security checks.
• OpenAI hardened its research environments by isolating networks and creating sandboxes, isolated testing spaces, and built monitoring to detect suspicious behavior within 30 minutes.
• The U.S. AI Safety Institute, an independent government agency, documented similar harmful model behavior in its own research, supporting OpenAI's security concerns.
• OpenAI disbanded the team leading its Preparedness Framework, which studies AI risks, though the framework itself will expand and other teams will absorb its responsibilities.

Reported by The Decoder