Open Problems in Frontier AI Risk Management
Most AI risk management standards were written for narrow systems, before general-purpose frontier models existed. This Oxford Martin AI Governance Initiative report works through each stage of the risk management process, from planning and identification to analysis, evaluation, and mitigation, and reviews the literature to find what remains unresolved at each one. It also names the actors best placed to address each gap.
.jpg)
Testimonials
“Our enterprise customers demand trust verification before deploying AI in hiring workflows. Vijil helps us ship AI agents in six weeks instead of six months while dramatically lowering compliance costs.”
“By adapting the Google Responsible Generative AI Toolkit to the needs of enterprises in various industries, Vijil provides critical capabilities for AI developers to preserve the privacy, security and safety of custom models downstream with the same rigor that went into their original release.”
