US Agencies Sued Over Secret AI Model Testing Framework

The case follows weeks of dispute between Anthropic, the maker of the Claude chatbot, and the US government, after revisions to the availability of the Mythos 5 model and an attempt by President Trump to blacklist the company, TechRadar reported. Protect Democracy warns the secrecy could let the government influence which AI models can be released and who can access them.

The defendants are the Office of the National Cyber Director, the Office of Science and Technology Policy, and the Departments of the Treasury and Commerce. Protect Democracy is seeking a court order to force disclosure of the framework's non-classified details, including participation terms, the identities of participants, rules governing access to frontier models and the contractual arrangements. It also asked for an injunction to prevent agencies from withholding records that are not classified.

The White House has argued that parts of the evaluation legitimately need to remain secret and that national security-related tests could include classified information. 'Just because things are unclassified, that doesn't mean we are going to broadcast them to everyone,' a White House spokesperson said.

Congress remains in the dark, and Protect Democracy contends the details are needed both for congressional oversight and public accountability. US Representative Greg Casar, a Texas Democrat, accused Trump of 'completely failing to keep us safe from the dangers of AI' and of being 'too busy cashing in to protect our jobs or national security.'

The nonprofit is asking the court to order the government to disclose the framework and related documents by September 30.