United States

Trump's wrath against Anthropic for refusing to use AI for mass surveillance: no agency will use it again

Start-up rejects unlimited access demanded by Pentagon, maintains ban on mass surveillance and autonomous weapons

27 February 2026aggiornato ore 8:50

2' min read

Translated by AI

Versione italiana

2' min read

Translated by AI

Versione italiana

Donald Trump steps into the clash between Anthropic and the Pentagon and announces that he has ordered 'all federal agencies to stop using' its artificial intelligence model. "Those left-wing human cases have made a disastrous mistake," the US president wrote on Truth.

The Battle

Anthropic 'cannot in good conscience' give in to the Pentagon's demands for unrestricted access to its technology. The artificial intelligence startup thus rejected the Pentagon's latest offer to defuse the impasse over the use of Claude, its chatbot, for military purposes. "Threats don't change our position: we cannot in good conscience" give in to demands, said CEO Dario Amodei about the ultimatum from the Pentagon.

Claude is the only AI model available in the military's most classified systems and the one with the best performance in the most sensitive intelligence activities. The Pentagon demanded unrestricted access but Anthropic refused to lift its security measures completely.

The start-up had been open to withdrawing them, however, safeguarding two areas: mass surveillance of Americans and the development of autonomous weapons, i.e. capable of striking without human intervention. The Pentagon had demanded more and a heated confrontation ensued.

In recent days, the defence minister Pete Hegseth had imposed an ultimatum on Anthropic, giving it until Friday to accept or reject his proposal.

The scenarios

If, as happened, Anthropic did not abandon its terms, the Department of Defence had pledged to declare the company a supply chain risk, a move that would prevent it from collaborating with other defence contractors.

The Pentagon also threatened to invoke the Cold War-era Defense Production Act to use Anthropic's software despite the company's objections.

Ft: Pentagon wants to use Ia to detect China vulnerabilities

The Pentagon aims to develop new artificial intelligence tools to automatically gather information on Chinese infrastructure and detect vulnerabilities in adversary electronic systems, with possible applications in military planning and cyber operations. The Financial Times reports this, citing informed sources.

According to the newspaper, the project aims to create Ia systems capable of analysing software and critical networks to identify exploitable weaknesses in case of conflict, facilitating targeted infiltration or disruption.

One of the priorities, reports one source, could be the identification of energy nodes - such as power plants that feed data centres - to compromise adversary AI-based systems.

It is not yet clear which companies will participate in the initiative. However, the issue fits into the ongoing clash between the Department of Defence and Anthropic over access to advanced AI models.

Brand connect

I prossimi eventi

Tutti gli eventi

Notizie e approfondimenti sugli avvenimenti politici, economici e finanziari.

World