The confrontation between Anthropic and the United States government escalated to unprecedented levels this week. President Donald Trump announced a ban on all Anthropic products across the federal government after the company refused to allow unrestricted military use of its AI model Claude. Hours later, OpenAI struck a deal with the Pentagon to fill the gap.
What happened
Defense Secretary Pete Hegseth gave Anthropic an ultimatum on Monday, February 24: open its AI technology for unrestricted military use by Friday, or face consequences. Anthropic refused, with CEO Dario Amodei stating the company "cannot in good conscience" agree to the Pentagon demands.
Trump responded by announcing on Truth Social a government-wide ban on Anthropic, calling it a "threat to national security." Hegseth went further: he threatened to invoke the Defense Production Act and designate Anthropic as a "supply chain risk" — a label normally reserved for U.S. adversaries like Chinese companies.
Why Anthropic refused
According to reports from Axios and Breaking Defense, Amodei's concerns include:
- Autonomous armed drones: AI making lethal decisions without human oversight
- Mass surveillance: AI-assisted tracking of dissent and citizen monitoring at scale
- No guardrails: the Pentagon wanted access with zero ethical restrictions
Anthropic's acceptable use policy prohibits using Claude to "directly harm people" and for "mass surveillance without consent." The Pentagon considered these restrictions unacceptable for defense operations.
OpenAI seized the opportunity
Just hours after Trump's announcement, OpenAI signed a deal with the Pentagon to deploy its models on the Department of Defense's classified network. The company said it would include "safeguards," though it did not specify what those entail.
In my experience covering the AI industry, this move from OpenAI is not surprising. The company quietly removed its "no military use" clause back in January 2024, laying the groundwork for exactly this kind of contract.
Anthropic vs OpenAI: military AI comparison
| Aspect | Anthropic | OpenAI |
|---|---|---|
| Military use stance | Refuses unrestricted use | Accepted Pentagon contract |
| Harm policy | Prohibits direct harm to people | Removed "no military" clause in 2024 |
| Government relationship | Banned from federal government | Active DoD contract |
| Legal response | Plans to challenge designation | Full cooperation |
| Defense Production Act | Threatened with invocation | Not applicable |
What is the Defense Production Act and why it matters
The Defense Production Act (DPA) is a 1950 law that allows the president to compel private companies to produce goods or services for national defense. It was used during COVID-19 to force production of ventilators and masks.
If invoked against Anthropic, it would be the first time the DPA has been used against an American tech company for refusing to modify the ethical restrictions of its software. This would set a dangerous precedent for the entire technology industry.
How this affects you
If you use Claude (Anthropic's chatbot), there is no direct impact on consumer service right now. The ban applies only to government use. However, the long-term implications could include:
- Regulatory pressure on all AI companies to cooperate with government demands
- Potential market fragmentation between "cooperative" and "resistant" AI companies
- Renewed debate about the ethical limits of artificial intelligence
Troubleshooting common concerns
Will Claude stop working?
No. The ban is for government use only. Claude remains fully available for consumers and private companies at claude.ai with no changes.
Does this affect Claude Code or the Anthropic API?
Only if you work directly on U.S. government contracts. For independent developers and private companies, API access remains unchanged.
Is OpenAI now the only option for government AI?
Not necessarily. Google (Gemini), Meta (Llama), and other providers also compete for government contracts. OpenAI was simply the first to capitalize on the situation.
What comes next
Anthropic announced on Friday that it plans to legally challenge the "supply chain risk" designation. After testing both platforms extensively, I believe this situation will reveal a lot about where the AI industry draws the line when the government comes knocking.
