Working with US CAISI and UK AISI to build more secure AI systems

OpenAI Blog · 10 months ago

OpenAI is collaborating with the US Cyber Incident Investigation and Supervision Instrument (CAISI) and the UK Artificial Intelligence Safety Institute (AISI) to improve AI system security. The partnership focuses on developing standards and practices for safer AI development, with both institutions contributing expertise in AI safety evaluation and risk assessment. This coordination aims to establish shared frameworks for identifying and mitigating security vulnerabilities across AI systems.