AI's Safety Report Card: None of the Big Five Pass as Agent Accountability Gap Widens

A nonprofit study just handed the AI industry its worst collective grade yet. Guidelight AI Standards evaluated Anthropic, OpenAI, Google, xAI, and Meta across six basic safety practices — logging, monitoring, gated actions, emergency shutdown capability, and more — applied to their own internal AI systems. The result: none of them fully pass. Anthropic and OpenAI tied for the highest mark at a limp C+. Google scraped a D+. xAI landed on D-. And Meta? A flat...