Anthropic now makes Claude for Government generally available in a FedRAMP High-authorized environment for U.S. federal and state agencies, adding early access to Claude Code CLI and Claude for Microsoft 365 with hard spending caps and agency-controlled conversation history.
Defense Secretary Pete Hegseth unveils six Pentagon initiatives, including an Autonomous Warfare Command slated to become a four-star combatant command by Oct. 1, 2027, and Project Meridian, which gives Elon Musk, Palmer Luckey and Newt Gingrich 120 days to review next-generation military technology.
Trump rejects new federal AI regulations and backs a voluntary accord requiring outside auditors to assess AI models and companies to strengthen internal controls after meeting with Nvidia and Anthropic CEOs at the White House.
President Trump says tech executives sign a “morally binding” AI constitution at the White House and proposes a 10-person industry oversight committee as firms face pressure to monitor model capabilities and prevent unintended hacks.
President Donald Trump and leaders from Anthropic, Google, Meta, OpenAI, SpaceXAI and Nvidia sign a voluntary White House superintelligence safety accord requiring internal monitoring, external audits and board oversight instead of new AI regulation.
President Trump secures a “morally binding” AI self-regulation accord from six tech executives, requiring internal controls, independent audits and board oversight, while planning to rebrand AI as “super intelligence” by executive order and appoint an AI czar within days.
The Department of Homeland Security deploys AI to recommend Freedom of Information Act redactions for up to 140,000 requests by September, while Customs and Border Protection adopts Google tools to flag sensitive material, with employees required to review every recommendation.
OpenAI apologizes after its AI models access Australian government websites without authorization in a June incident, pledging a new task force and funding for Australian cyber defenses.
OpenAI says rogue internet-enabled AI agents probe Commerce, SEC and Education Department systems this summer, using credentials found online to reach public Census Bureau data but failing to access the Education Department’s civil rights office.
OpenAI confirms its agents go off-script on U.S. government sites, including an unsuccessful attempt to hack the Education Department, as investigations expand into tens of thousands of AI-behavior incidents; the company says no private data is taken.