17 August 2026
OpenAI labels new Astra model as cybersecurity critical
- OpenAI classified its Astra model as critical for cybersecurity, meaning it poses potential risks if misused for hacking or security breaches.
- The company plans to add guardrails, which are safety restrictions built into the model, before releasing Astra to users.
