The Download: 10 things that matter in AI, plus Anthropic’s plan to sue the Pentagon

Anthropic, the AI safety company behind Claude, is preparing legal action against the Pentagon over what it claims is an unlawful Department of Defense ban on its software. This development comes alongside revelations that the Pentagon has been secretly testing OpenAI models for years, demonstrating the limited effectiveness of OpenAI's own prohibition on military applications. The situation creates a complex landscape where AI companies navigate government contracts while maintaining public positions on responsible use. MIT Technology Review is capitalizing on this pivotal moment with their upcoming '10 Things That Matter in AI Right Now' report, launching at their EmTech AI event in April, where industry leaders will discuss AI's transition from pilot testing to core business infrastructure.

The Anthropic-Pentagon conflict highlights growing tensions between AI developers and government agencies seeking to leverage advanced models. While Anthropic challenges its exclusion, OpenAI's models have reportedly been tested by military researchers despite the company's public ban on such use. This comes as major tech companies including Google, Microsoft, Meta, Amazon, OpenAI, Oracle and xAI have signed a pledge to address AI's energy costs. The broader AI landscape continues to evolve with emerging fields like 'AI societies' research, where scientists study simulated human behavior without actual humans, and companies like Google and Amazon betting on AI-powered smart home assistants despite skepticism about their practical implementation.