Artificial intelligence lab Anthropic has no intention of easing its usage restrictions for military purposes, a person familiar with the matter said on Tuesday. Talks continue after a meeting to discuss its future with the Pentagon. The dispute centers on Anthropic military restrictions that prevent technology use for autonomous weapons targeting and domestic surveillance.
The meeting between Anthropic CEO Dario Amodei and US Defense Secretary Pete Hegseth was scheduled to hash out a months-long conflict. The AI startup has refused to remove safeguards limiting how the Pentagon can deploy its technology. Pentagon officials have argued the government should only be required to comply with US law, not company policies.
Pentagon Issues Ultimatum
During the meeting, Hegseth delivered an ultimatum to Anthropic. The choice was clear: get on board or the government would take drastic action, people familiar with the matter said. The options included labeling Anthropic as a supply-chain risk or having the Pentagon invoke the Defense Production Act. This law would force Anthropic to change its rules under government compulsion.
The government gave Anthropic until Friday at 5 p.m. to respond, according to a senior Pentagon official with knowledge of the matter. This deadline adds urgency to negotiations that have dragged on for months. The outcome will set precedent for how AI companies interact with military customers.
The Pentagon did not immediately respond to a comment request. An Anthropic spokesperson said Tuesday’s meeting continued good-faith conversations about usage policy. The company aims to support the government’s national security mission in line with what models can reliably and responsibly do. This statement suggests willingness to find common ground while maintaining core principles.
Military AI Contracts at Stake
The Pentagon has been negotiating AI contracts with multiple large language model providers. These include Alphabet’s Google, xAI and OpenAI. The contracts will shape the future of military use of artificial intelligence for battlefield applications. Potential uses span autonomous drone swarms, robots and cyber attacks.
Until recently, Anthropic was the only LLM provider on classified networks. This exclusive position gave the company significant influence. However, this week the Pentagon announced it had reached an agreement with xAI to deploy it across classified networks. Reuters has previously reported that the Pentagon plans to move all AI companies to classified networks.
The expansion of providers reduces Anthropic’s leverage in negotiations. The Pentagon now has alternatives if Anthropic refuses to compromise. However, the confrontation raises broader questions about government power over technology companies. The Anthropic military restrictions dispute tests the boundaries of corporate autonomy.
Venezuela Raid Questions Intensified Dispute
The Pentagon’s fight with Anthropic reached a fever pitch earlier this month. The government grew concerned that the company had asked questions about how its AI products were used during the Venezuela military raid that captured President Nicolas Maduro. This inquiry suggested Anthropic was monitoring how its technology might be deployed in sensitive operations.
During the meeting with Hegseth, Amodei clarified the situation. He said Anthropic did not raise concerns to Palantir or the Pentagon about whether the company’s AI products were used during the Venezuela raid, the source said. This statement aimed to alleviate concerns about corporate oversight of military operations.
Amodei also said the safeguards currently in place would not pose a problem to the Defense Department’s current operations. This argument suggests the restrictions are not as restrictive as Pentagon officials believe. The gap between perception and reality may offer room for resolution.
Defense Production Act Threat
Hegseth presented two stark options during the meeting. The Pentagon would either invoke the Defense Production Act to compel Anthropic to comply with its demands. Alternatively, it would deem the company a supply chain risk, a determination typically imposed on companies from foreign adversaries. This could upend Anthropic’s business with other companies that do business with the US government.
The Defense Production Act grants the president broad authority to influence domestic industry for national defense purposes. Invoking it against a private company over contractual terms would be unusual. It would signal extraordinary government interest in accessing Anthropic’s technology without restrictions.
Franklin Turner, a government contracts lawyer at McCarter & English, described the situation as unprecedented. “This specific scenario is unprecedented and will almost certainly trigger a raft of downstream litigation if the Administration takes adverse action against Anthropic here,” he said. Legal challenges would likely follow any punitive measures.
Supply Chain Risk Designation Consequences
Designating Anthropic as a supply chain risk carries severe implications. This label is typically reserved for companies linked to foreign adversaries. It would effectively bar Anthropic from doing business with government agencies and contractors. Other companies might avoid partnerships to maintain their own government relationships.
The supply chain risk designation would transform Anthropic’s business landscape overnight. The company’s valuation and growth prospects depend partly on government and corporate contracts. Losing access to these markets would fundamentally alter its trajectory.
The Pentagon’s willingness to consider this option demonstrates frustration with Anthropic’s position. It also reflects growing government expectations that technology companies will prioritize national security over corporate policies. This tension will likely intensify as AI capabilities advance.
Ethical AI and Military Applications
Anthropic has positioned itself as an AI company committed to safety and ethical deployment. Its usage policies reflect concerns about how technology might cause harm. Autonomous weapons and domestic surveillance represent precisely the applications many AI ethicists seek to prevent.
The company faces pressure from employees and the AI ethics community to maintain strong safeguards. Backing down to Pentagon demands could trigger internal dissent and public criticism. This stakeholder dynamic complicates negotiations beyond simple business considerations.
However, the Pentagon argues that national security requires access to cutting-edge technology. Adversaries will not abide by ethical restrictions. The government must have the best tools to protect American interests. This fundamental tension between corporate ethics and national security drives the current dispute.
Industry Implications
The outcome of this dispute will affect the entire AI industry. Other companies watching closely will adjust their own policies and negotiating positions. A government victory could encourage the Pentagon to pressure other providers. A company victory might embolden others to maintain strict usage policies.
The Pentagon’s agreement with xAI this week signals its determination to diversify providers. It reduces dependence on any single company. This strategy strengthens the government’s position in negotiations with Anthropic and future partners.
The AI industry has largely cooperated with government requests in recent years. However, the line between cooperation and compromise remains contested. Anthropic military restrictions represent a boundary the company is unwilling to cross. Whether it can maintain that position against government pressure remains unclear.
Friday Deadline Approaches
Anthropic now faces a Friday 5 p.m. deadline to respond to Pentagon demands. The company must decide whether to modify its usage policies or accept consequences. Either choice carries significant implications for its future.
The meeting between Amodei and Hegseth continued good-faith conversations, according to the company. This suggests dialogue remains open despite the ultimatum. However, the fundamental disagreement over military restrictions persists. Bridging this gap will require creativity from both sides.
The dispute highlights growing pains as AI technology matures. Companies developed usage policies in relative autonomy. Now government customers demand access that challenges those policies. Resolving these conflicts will shape the future relationship between AI developers and national security institutions.
For now, Anthropic holds its position on military restrictions. The Pentagon presses its case with ultimatums and alternatives. Friday’s deadline will force resolution one way or another. The outcome will echo through the AI industry and government contracting for years to come.
