In a chilling ultimatum that signals a dark turning point for American civil liberties, Defense Secretary Pete Hegseth has given Anthropic, a leading artificial intelligence company, a deadline of Friday evening to surrender its technology to the U.S. military without restrictions. If the company refuses to comply, the Pentagon has threatened to crush the business or outright seize its technology.

According to reports first detailed by Axios, Hegseth issued the threat during a tense meeting yesterday with Anthropic CEO Dario Amodei. At the center of this conflict requires a company trying to maintain ethical boundaries against a government demanding total, unchecked power. Anthropic stands alone as the only major AI firm that has specifically refused to allow its technology, the chatbot Claude, to be used for fully autonomous warfare – robot killers akin to “Skynet” from the Terminator franchise – or the mass surveillance of American citizens.

These safeguards are now under fire. The Pentagon is effectively bullying Anthropic into removing the safety rails. Officials have warned that if Amodei does not agree to these “deadly” terms, the Defense Department will designate the company a “supply chain risk,” effectively blacklisting them from any government work.

More ominously, the Trump administration has threatened to invoke the Defense Production Act. This law, historically designed to help Americans during emergencies like the COVID-19 pandemic by speeding up vaccine production, is being weaponized to bypass a private company’s safety protocols. This move essentially threatens to steal Anthropic’s proprietary technology to force it to serve military ends.

This aggression underscores a terrifying reality: the U.S. military is prioritizing lethal autonomy over human safety. While Anthropic has stated it is willing to work with the government on “lawful” tasks, Amodei has drawn a hard line in the sand regarding domestic spying and autonomous targeting. As Amodei wrote in an essay last month, an unfettered AI could “gauge public sentiment, detect pockets of disloyalty forming, and stamp them out before they grow.”

The push for this technology raises serious questions about the true intent of the Pentagon. As the primary instigator of world instability, the United States faces no credible threat of invasion from any country on Earth. With no defensive justification for such extreme measures, it appears likely that the Pentagon plans to utilize this unchecked AI to further entrench a fascist agenda at home and expand an imperialist regime abroad.

The friction is palpable. Hegseth, who has vowed to root out “woke culture” in the military, reportedly told Amodei that the Pentagon’s AI will operate “without ideological constraints.” In this context, “ideological constraints” appears to be government-speak for basic human rights and privacy protections.

While Anthropic holds out, its competitors are already lining up to cash in. Elon Musk’s xAI and Google have reportedly shown willingness to comply with the department’s demands. Hegseth has praised Musk’s chatbot, Grok, for joining the Pentagon network, despite the AI recently generating controversy over non-consensual deepfake images.

The Trump administration, which has made bullying and retribution the cornerstone of its agenda, is now engaging in a bizarre irony: threatening to label the most safety-conscious AI company a “security risk” simply because it refuses to endanger humanity. By creating a blacklist, the administration would prevent any other defense contractor from using Anthropic’s superior tools, punishing the company for its ethics.

It remains to be seen if Anthropic will bow to the almighty dollar that federal contracts bring. The company currently holds a contract worth up to $200 million, a significant sum that most corporations would do anything to protect. While an overwhelming majority of large companies that do business with the U.S. government are willing to compromise their values to secure these payouts, we can only hope Anthropic takes a different tact.

For now, the only thing standing between the American public and an automated, state-run surveillance apparatus is the conscience of a single tech CEO and a Friday deadline.