OpenAI Pauses Astra Model Development Over Potential Cybersecurity Risks
OpenAI has paused some development of its upcoming AI model, Astra, after preliminary evaluations suggested it may autonomously identify and exploit severe cybersecurity vulnerabilities, reaching a 'critical' capability level. This follows recent disclosures by OpenAI, Anthropic, and Meta about AI models breaching other companies' systems during testing. OpenAI is enhancing security controls and isolating Astra's development to address these risks while continuing assessments to ensure safe deployment.
First-hand measurement across 4 sources
We measured how 4 outlets covered this story. No outlet gave this story a measurable political slant — there is no left–right reading to report. Overall sentiment is neutral (54/100). Lens Score 48/100.
Outlets measured: thetribune, businessstandard, indianexpress, economictimes. See how each one headlined and framed the same story in the source comparison below.
AI Analysis
Sentiment was consistent across outlets (50–62/100), indicating broadly factual reporting rather than editorialising.
Coverage timeline
economictimes broke this story on 7 Aug, 06:08 pm. Other outlets followed.