← signals
2026-08-11·OPENAI·security risk
meddown

OpenAI said it has paused some internal activities involving its unreleased Astra model after preliminary evaluations...

OpenAI said it has paused some internal activities involving its unreleased Astra model after preliminary evaluations could not rule out a 'Critical' capability level — meaning it might autonomously launch cyberattacks against sophisticated defenses (CNBC).

window 30devidence 59confidence score 100

confidence score

Strong evidence: 12 independent source classes support this read.

100
medium confidence12 independent source classesothernewsregionalmarketpasses publish gate

signal brief

OpenAI said it has paused some internal activities involving its unreleased Astra model after preliminary evaluations could not rule out a 'Critical' capability level — meaning it might autonomously launch cyberattacks against sophisticated defenses (CNBC). The company is tightening controls: isolated testing environments, universal monitoring for risky actions and misalignment across all agentic applications, and additional monitoring and detection (CNBC). The pause comes as U.S. lawmakers push the 'AI Kill Switch Act' to require AI companies to maintain the ability to shut down, throttle, or suspend models, and as the EU gains new powers to inspect models due for release and restrict market access (CNBC). In parallel, OpenAI's head of ethics, Chloé Bakalar, has departed after less than a year, following exits by the head of safety systems and a former head of mission alignment (FT). OpenAI also expanded its Daybreak cyber-defense program with the new GPT-5.6-Cyber model across two access tiers (TechCrunch; Axios). For AI-infrastructure watchers, any delay in Astra-class deployment pushes out a major wave of inference compute demand and adds regulatory overhang to data-center buildout.

What the sources said

  • "While we continue to benchmark and assess this model, our preliminary evaluations indicate strong enough performance that we cannot rule out Critical capability level at this time." — OpenAI statement via CNBC
  • "Chloé Bakalar's departure is one of several high-profile exits at the group in recent weeks as safety concerns mount." — FT
  • "As the threat landscape evolves, we're putting frontier intelligence in the hands of trusted defenders before attackers can deploy offensive AI at scale." — OpenAI via CNBC
  • "The cybersecurity world is rapidly changing—threat actors will increasingly use AI to conduct cyberattacks at unprecedented speed and scale, including in fully autonomous ways." — OpenAI via TechCrunch

source data used

Decision support, not stock advice. This signal is research with cited evidence — not a recommendation to buy, sell, or hold any security.