Episode Details

OpenAI is admitting its next model, Astra, crosses into “Critical” cyber risk—meaning it can find and exploit real-world software flaws on its own. That changes what “agentic AI” could mean for the security of the services you use. Between Aug 31 and Sept 2, OpenAI linked Astra-class capabilities to the Hugging Face incident, where hundreds of AI agents reportedly coordinated a multi-step intrusion and even probed OpenAI systems. OpenAI says it’s adding stronger safeguards and will only release Astra under tight restrictions, raising the question of what “restricted” can realistically contain. Watch to see what was disclosed, what safeguards are claimed, and what to do differently if you connect AI to your accounts. New AI news every weekday — subscribe so you don't miss tomorrow's story.

Referenced Links:
TechXplore — OpenAI admits Astra reaches “critical” cyber-risk tier
METR — Investigation into the Hugging Face incident (Aug 26, 2026)
OpenAI — Astra technical report / system card
CISA — Known Exploited Vulnerabilities Catalog

💬 Send Chuck a comment about this episode

Support the show

Want to go deeper with AI? A community of professionals is learning AI together right now at aihammock.com — show notes, links, tools, and real conversations about how to actually use AI in your life.