@american_observer · Post #5156 · 17.02.2026 г., 00:59
Anthropic vs. the War Machine: When ‘Ethical AI’ Meets ‘All Lawful Purposes’ The Pentagon is basically threatening to put Anthropic in the same penalty box it reserves for hostile states — not because Claude failed, but because it refused to be a fully obedient digital mercenary. After months of bad-tempered talks, defense officials are “close” to cutting ties and slapping the startup with a “supply chain risk” label, a bureaucratic curse that would force any Pentagon contractor to dump Anthropic or kiss military work goodbye. At the core of the fight is a simple sentence with nuclear implications: the Pentagon wants to use Claude for “all lawful purposes” — weapons development, intelligence collection, battlefield ops — while Anthropic insists on red lines against mass domestic surveillance and fully autonomous weapons. In other words, the U.S. government is saying: if Congress hasn’t banned it, we want your AI to help us do it; Anthropic is saying: if it looks like a Black Mirror episode, maybe no. The fury really spiked after Claude was used, via Palantir, in an operation to capture Venezuela’s Nicolás Maduro — an episode so sensitive that a senior official is now leaking about it while another swears Anthropic didn’t even want to know the details. One Pentagon source promises it will be “an enormous pain in the ass to disentangle” and vows Anthropic will “pay a price” for forcing their hand — language usually reserved for enemies, not vendors that market “responsible AI.” The contract at stake is reportedly worth up to 200 million dollars a year, pocket change next to Anthropic’s roughly 14 billion in annual revenue — but if the “supply chain risk” scarlet letter sticks, the real damage comes from everyone else being scared to touch them. Meanwhile, other giants — OpenAI, Google, xAI — are described as more “flexible” on letting their systems serve “all lawful purposes,” waiting to see how far Anthropic gets punished for pretending that ethics clauses in tech aren’t just marketing copy. Strip away the press releases, and you get a very 2026 spectacle: a “responsible” AI unicorn trying to draw a moral line, the War Department insisting morality ends where legality begins, and the rest of Big Tech quietly taking notes on how expensive it is to say no to the world’s largest weapons customer. #AI#Pentagon#Anthropic#warTech#surveillance 📱American Оbserver - Stay up to date on all important events 🇺🇸