Anthropic just had a falling out with its biggest potential customer.
The artificial intelligence company behind Crowder filed a lawsuit in the U.S. District Court for the Northern District of California on Monday, naming the Treasury Department, Commerce Department, State Department, Health and Human Services Department, Veterans Affairs Department, General Services Administration and several other federal agencies as defendants.
Anthropic said the U.S. government effectively blacklisted its AI systems for federal procurement and failed to follow any of the legal procedures required to actually ban the vendor.
There was a lack of formal decisions, inter-agency reviews, documentary evidence and an evaluation of less restrictive alternatives such as conditional approval or safety audits, the report said.
Officials justified the restrictions internally on national security and supply chain grounds, and then allowed the directive to spread informally through centralized procurement channels until Anthropic was excluded from federal contracts across the board, according to the complaint.
The timing makes this more than just a procurement dispute.
The U.S. government is in the midst of the largest AI adoption push in federal history, using OpenAI’s ChatGPT as its tool of choice. Agencies are deploying generative AI for cybersecurity, intelligence analysis, management automation and internal decision-making. These contracts are large, multi-year, and increasingly important to how government operates.
Being excluded from this market is not a minor business setback for any AI company that wants to be taken seriously at an institutional level, but an existential competitive issue.
Anthropic is asking the court to declare the restrictions illegal and prevent agencies from enforcing them. If successful, the ruling would reopen federal procurement and could set a precedent for how far agencies can go to restrict AI vendors on national security grounds without complying with their own rules.
The government has not responded publicly to the filing, but an Axios report on Tuesday cited people familiar with the matter as saying that the White House is preparing an executive order to formally direct the federal government to remove Anthropic’s artificial intelligence from its operations.
