“There Is No Distinction Between Offensive and Defensive”: The Emails That Reveal America’s Own AI Weapons Ambitions

Newly unsealed court documents show the Pentagon punished an American AI company for refusing to allow autonomous weapons and mass surveillance. The West's favourite accusation against China now points inward

The Line That Changes the Story

For weeks, watchchina.ai has documented the West's central accusation against Chinese AI: that Beijing is building "AI-enabled techno-authoritarianism" — autonomous weapons, mass surveillance, and military AI without ethical restraint. That accusation deserves scrutiny. But intellectual honesty requires applying the same scrutiny to the accuser — and newly unsealed court documents have made that impossible to avoid.

Court documents unsealed July 2, 2026, in the U.S. District Court for the Northern District of California reveal a tense private exchange between Anthropic CEO Dario Amodei and Emil Michael, the Pentagon's Under Secretary of Defense for Research and Engineering. The emails establish that the core dispute was never about capabilities — it was about control, specifically whether the Pentagon could deploy Anthropic's Claude model for fully autonomous weapons and domestic surveillance programs without restriction.

When Amodei tried to preserve a distinction between offensive and defensive military uses, Michael's reply was categorical: "There is no distinction in our world between weapons that are defensive or offensive." Anthropic's guardrails were left with, in the reporting's words, no foothold.

An American Company, Punished for Its Red Lines

The consequence of Anthropic holding those red lines is the part that should reframe the entire Western narrative about who, exactly, is building authoritarian AI.

Anthropic had made clear since signing its Pentagon contract in 2025 that it did not want its technology used for mass surveillance of people in the United States or for fully autonomous weapons systems. When it refused to remove those two restrictions, the Pentagon designated Anthropic a "supply chain risk to national security" — a label previously reserved for companies associated with foreign adversaries — barred defence contractors from working with it, and began removing Claude from military systems on a 180-day timeline.

Read that sequence again. An American AI company was formally branded with the same designation used for hostile foreign entities — not for a security failure, but for refusing to let its AI be used in autonomous weapons and the mass surveillance of American citizens. As watchchina.ai has repeatedly argued, the picture is never as simple as "democratic West versus authoritarian China." Here is the plainest possible evidence: the pressure to build unrestricted military and surveillance AI is not a uniquely Chinese phenomenon. It is a structural feature of great-power competition — and it is operating, right now, inside the United States.

The Uncomfortable Symmetry — and a Detail That Demands Scrutiny

watchchina.ai does not draw a false equivalence here, and the differences matter. The United States has independent courts — a federal judge blocked the Pentagon's designation, something no Chinese company could obtain against the PLA.

Anthropic could sue the government publicly, and nearly forty employees from Google and OpenAI filed briefs supporting its stand. That capacity for internal dissent and legal challenge is a genuine structural difference between the two systems, and it should not be flattened away.

But the emails reveal something that complicates any triumphant reading of that difference. Emil Michael — the official who pressed hardest for Anthropic to remove its guardrails — held between $2 million and $10 million in stock in Perplexity, a direct Anthropic competitor, and had earlier sold xAI stock, another competitor, for a reported return of up to 4,800%. The man demanding an American AI company drop its safeguards against autonomous weapons stood to benefit financially from that company's competitors. When watchchina.ai examines China's fusion of state, military, and commercial interest — the defining feature of its AI-industrial complex — this is a reminder that the conflicts of interest run in every direction, and that the moral clarity the West claims in the AI race is, on close inspection, considerably murkier than the official narrative allows.

The right posture toward both superpowers is the same: verify, don't trust — and watch what they do, not what they say.

Sources: Wall Street Journal, Gizmodo, Tech Times, Congressional Research Service, Electronic Frontier Foundation, Atlantic Council, ProPublica / The Lever, Wikipedia