February 26, 2026
AI giant – Anthropic — just told Secretary of Defense Pete Hegesth – thanks but no thanks….you can’t use it’s AI model Claude that way.
Anthropic CEO Dario Amodei just posted this lengthy statement on the company website saying that they are not going to allow Claude to be used in the way Hegseth has demanded. Hegseth gave Anthropic a deadline of Friday to either comply with his demand of expanded use of Claude throughout the US military or basically be stripped of their government contracts and be labeled a “risk” to the supply chain…which is a fancy way of saying blacklisted by the US government. Just a short time ago Anthropic said bet. We’re still not going to do it.
So – here’s what’s happening. Anthropic was awarded a $200-million contract with the pentagon in July to develop AI capabilities for the US military. It was reported that Anthropic’s AI model Claude was used by the US military in January to model the Venezuela raid to capture Nicolas Maduro. A few days after that raid – obviously riding a successful high — the department of defense-slash-war put out this press release saying they are going to expand their use of AI as part of their overall military strategy. Now – Anthropic has been on board with the pentagon so far. They were the first company to deploy their AI model on the government’s classified networks and they say Claude is “extensively deployed” already across the department of war and other national security agencies for intelligence analysis, modeling and simulation, operational planning, cyber operations and more.
But here’s the thing – Anthropic has some company guardrails when it comes to the use of it’s AI models. Amodei says “in a narrow set of cases, we believe ai can undermine rather than defend democratic values. Some uses are simply outside the bounds of what today’s technology can safely and reliably do. Anthropic has two red lines — they will not allow their AI model to be used for what the company calls “mass domestic surveillance” meaning mass AI-driven spying on you and me. Amodei says the current laws on the books that protect our privacy are so old and AI is so powerful that right now – it can legally find ways access your movements, web browsing and more without a warrant on an automatic and massive scale. So Claude driven mass domestic surveillance – red line. And the other red line is on what the company calls fully autonomous weapons…meaning weapons that aim and fire on a target without a human behind it. Anthropic says the technology is good but it’s not good enough at this point and Amodei says “we will not knowingly provide a product that puts America’s warfighters and civilians at risk.”
Well earlier this week Secretary Hegseth met with Amodei about these red lines and reiterated what the pentagon demanded in this letter from January which said if you want to keep your contract with the pentagon you must remove those safeguards – because we’re still operating within the law — and we want unrestricted use of you’re AI models. Oh – and Hegseth threatened that if Anthropic declines – he will get President Trump to invoke the defense production act and force Anthropic to comply.
Today – Amodei said “these threats do not change our position: we cannot in good conscience accede to their request.” We’ll see how the department of defense responds.
Anthropic’s statement here.
Department of War’s plan to expand AI here.
