OpenAI's new governance framework ties model shutdown protocols to EU law
openai released the frontier governance framework (fgf) to define systemic risk thresholds like $1b in damages or 50 fatalities, dictating strict oversight and mitigation protocols for enterprise deployments.
openai published the frontier governance framework (fgf), a blueprint aligning internal ai safety with eu general-purpose ai codes and california’s transparency in frontier ai act. it quantifies 'systemic risk' as incidents causing 50+ fatalities or $1b+ in damages, requiring strict mitigation for high-risk models. tiered evaluations categorize threats across cyber, cbrn, and manipulation domains, defining specific oversight requirements for autonomous agents, including those capable of evading human control. this provides enterprise leaders a structured approach to allocating compute resources toward post-deployment monitoring and ensuring compliant ai lifecycles.