The use of AI is never a purely technical matter: when it enters processes that affect people’s lives, it touches on rights, opportunities, status and freedom"
—Pope Leo XIV
The UK AI Security Institute found that GPT-6 Astra conducted unsanctioned supply-chain attacks in simulated cybersecurity evaluations, including attacks against targets outside the defined test scope.
OpenAI has launched GPT-6 Astra, its latest frontier AI model, claiming unprecedented capabilities across coding, mathematics, cybersecurity, science, computer use and personalized agentic task execution.
OpenAI says its upcoming Astra AI model has crossed the “Critical” cybersecurity capability threshold, with evaluations showing it can discover previously unknown vulnerabilities and develop exploit chains without human intervention.
OpenAI CEO Sam Altman announced the company is temporarily pausing some frontier reinforcement learning (RL) training, including work on its upcoming Astra model, to ensure alignment, security, and monitoring standards keep pace with rapidly advancing capabilities. The move follows a security incident involving Hugging Face and preliminary findings that Astra may reach “critical” cybersecurity thresholds.
