In March 2026, the Trump administration labeled Anthropic, a company founded in 2021 known for its commitment to AI safety, as a supply chain risk, leading to the federal government ceasing its use of Anthropic’s large language model, Claude. This drastic decision came after Anthropic refused to remove key safeguards against domestic surveillance and autonomous weapons from its products supplied to the Pentagon. The action underscored a broader pattern in which authoritarian governments, like the Trump administration, undermine existing safety standards, framing them as ideological barriers rather than technical necessities. This strategy affects competitiveness in the defense sector, where lucrative contracts create pressure on companies to compromise ethical protections for access to critical government resources.

OpenAI: OpenAI is a major AI developer known for its GPT series of large language models and its work on generative AI applications. It has competed directly with Anthropic since the latter’s founding by former OpenAI researchers. In the events described, OpenAI quickly secured the Pentagon supply role vacated by Anthropic.
Anthropic: Anthropic is an AI company founded in 2021 by researchers focused on building advanced language models with strong safety measures. It has distinguished itself through safeguards against uses like domestic surveillance and autonomous weapons in its Claude model. In March 2026, the company lost Pentagon contracts after refusing to alter those safeguards at the Trump administration’s direction.
Donald Trump: Donald Trump is the President of the United States. He directed the federal government to cease using Anthropic’s Claude model, citing national security concerns tied to the company’s safety policies. The action aligned with broader administration efforts to reshape AI ethics standards.
Brennan Center: The Brennan Center is a legal policy and advocacy organization that examines issues at the intersection of technology, democracy, and civil rights. It has tracked how government contract processes are used to redefine AI ethics standards. Its documentation highlights pressures on companies maintaining certain civil rights protections.

`json
{
“Policy Pressure”: “The Preventing Woke AI executive order has been applied to frame standard safety and ethics features as ideological obstacles in federal AI dealings.”,
“Governance Pattern”: “Authoritarian approaches to AI often target voluntary or self-imposed safety policies rather than relying solely on outright censorship.”,
“Contract Incentives”: “Large defense AI contracts create competitive dynamics where companies risk losing access to valuable data and future work by maintaining independent guardrails.”
}
`