#ai-safety-standards

#ai-safety-standards

The US Department of Commerce's CAISI division signed agreements with Google DeepMind, Microsoft, and xAI to vet AI models before public release, marking a shift toward proactive government-led security testing of advanced AI systems.

UL Solutions launches its first AI product certification standard (UL 3115) to ensure AI-embedded products are safe, robust, and human-controlled, addressing the lack of government oversight in rapidly evolving AI technology.

A House amendment to prevent government blacklisting of firms refusing to deploy high-risk AI technology failed 16-25, after the Pentagon threatened to blacklist Anthropic for maintaining AI safety standards.