OpenAI to work with Pentagon after Anthropic dropped by Trump over company’s ethics concerns
OpenAI announced a contract with the U.S. Pentagon that includes strict prohibitions on mass surveillance and autonomous lethal weapons, a stance that contrasts with a recent Trump directive to cease using rival Anthropic's AI. The deal underscores growing tensions between government demand for adv…
OpenAI disclosed on Friday night that it has entered into a classified agreement with the U.S. Department of Defense to provide its generative‑AI models for secure military networks. The contract, which the company says incorporates explicit safeguards against domestic mass surveillance and fully autonomous weapon systems, marks a rapid shift in the AI‑defense landscape after President Donald Trump ordered a blanket stop to the use of a competitor’s technology.
Background: Trump’s crackdown on Anthropic
Earlier in the week, the Trump administration issued an executive directive instructing all federal agencies to "IMMEDIATELY CEASE" any deployment of Anthropic’s Claude model. Anthropic, a San Francisco‑based AI startup known for its safety‑first branding, had been in protracted negotiations with the Pentagon. Defense officials wanted broader access to Claude’s capabilities for national‑security tasks, while Anthropic insisted on two non‑negotiable red lines: no use for mass domestic surveillance and no integration into autonomous lethal weapons.
When the Pentagon pressed Anthropic to relax those constraints, the company responded that its ethical stance would not change, stating that the two exceptions had never impacted a single government mission. Trump, reacting on his Truth Social platform, accused Anthropic of “strong‑arming” the Pentagon and demanded compliance with what he framed as constitutional priorities.
OpenAI’s response and the new Pentagon deal
Within hours of the presidential order, OpenAI CEO Sam Altman announced that his firm had secured a deal with the Pentagon that mirrors Anthropic’s safety concerns. In a post on X, Altman wrote that two of OpenAI’s core safety principles are a prohibition on domestic mass surveillance and a requirement that humans retain responsibility for any use of force, including autonomous weapon systems. He claimed the Pentagon “agrees with these principles, reflects them in law and policy, and we put them into our agreement.”
Altman further suggested that the agreement could serve as a template for future contracts, urging the Department of Defense to offer the same terms to all AI providers as a way to “de‑escalate away from legal and governmental actions and toward reasonable agreements.” The deal reportedly allows OpenAI’s models to operate in classified environments, provided they are not employed for unlawful activities such as domestic surveillance or autonomous offensive weapons.
Industry reaction and internal dynamics
The development has sparked a mixed response among AI workers and competitors. Nearly 500 employees from OpenAI and Google signed an open letter declaring “we will not be divided,” warning that the Pentagon’s pressure could pit companies against each other. The letter argues that the defense department is attempting to force a concession from one firm while threatening the other, thereby creating a “divide‑and‑conquer” scenario.
Altman addressed OpenAI staff in an internal memo obtained by Axios, emphasizing that the issue now extends beyond a single rivalry and concerns the entire industry’s ethical framework. He reiterated the company’s long‑standing red lines and noted that the contract would explicitly exclude any use that violates those boundaries. The memo also highlighted OpenAI’s ongoing $110 billion funding round, which would value the company at roughly $840 billion, underscoring the high stakes of maintaining a responsible public image while courting lucrative government business.
What the Pentagon gains and what remains uncertain
From the defense perspective, access to OpenAI’s large‑scale language models could enhance intelligence analysis, logistics planning, and rapid decision‑making in classified settings. However, the agreement’s safety clauses leave open questions about enforcement, oversight, and how “human‑in‑the‑loop” requirements will be operationalized in real‑time missions.
Critics argue that even with contractual language, the line between permissible and prohibited uses can blur, especially when AI tools are embedded in broader systems. The Pentagon has not disclosed the exact scope of the deployment, citing national‑security concerns, so analysts will watch for future disclosures or whistle‑blower reports that could clarify compliance.
Looking ahead: broader implications for AI governance
The OpenAI‑Pentagon deal may set a precedent for how the U.S. government engages with private AI firms. If the safety provisions hold, they could become a de‑facto standard for future contracts, potentially easing industry fears about being forced to compromise core ethical commitments. Conversely, if enforcement proves weak, the agreement could embolden other agencies to push for fewer restrictions, reigniting the Anthropic‑style standoff.
Stakeholders—including civil‑rights groups, tech ethicists, and congressional oversight committees—are likely to scrutinize the contract’s language and its real‑world application. The outcome could influence upcoming legislation on AI use in defense, as well as shape public perception of how AI companies balance profit, innovation, and moral responsibility.
Why it matters
The deal highlights the clash between national‑security demands for powerful AI and industry‑wide ethical standards that aim to prevent misuse in surveillance and autonomous weapons.
Key points
- OpenAI secured a Pentagon contract that bans domestic mass surveillance and autonomous lethal weapons
- Trump ordered a halt to Anthropic’s AI use after the company refused to relax its safety safeguards
- Anthropic’s red lines remain unchanged, prompting a broader industry debate on AI ethics
- OpenAI’s agreement could become a template for future government‑AI contracts
- The deal comes amid OpenAI’s $110 billion funding round, valuing it at $840 billion
Frequently asked questions
What safety guarantees does OpenAI’s Pentagon deal include?
The contract explicitly prohibits the use of OpenAI’s models for domestic mass surveillance and for fully autonomous weapons that can kill without human intervention, and requires human oversight for any force‑related decisions.
Why did President Trump order a stop to Anthropic’s AI use?
Trump’s directive followed a breakdown in negotiations where Anthropic refused to loosen its ethical restrictions, prompting the administration to demand compliance or face consequences.
How might this agreement affect other AI companies?
If the Pentagon adopts OpenAI’s safety terms as a standard, other AI firms may be asked to sign similar contracts, potentially shaping industry‑wide norms for government collaborations.
What is the significance of OpenAI’s $110 billion funding round?
The massive fundraising round values OpenAI at $840 billion, underscoring the company’s financial clout and the strategic importance of securing government contracts while maintaining ethical credibility.
