AI-Powered Regime Change: Old Wine, New Algorithm
The Wall Street Journal recently let slip a peculiar detail regarding the US military's 2026 incursion into Venezuela: the deployment of Anthropic's AI model, Claude, in an audacious attempt to abduct President Nicolás Maduro. This operation, which reportedly inflicted upon Caracas a bombing campaign and claimed 83 lives, stands in stark contrast to Anthropic's own terms of service, which
expressly forbid using Claude for violent ends, weapons development, or surveillance. One might reasonably wonder how a 'prohibition' dissolves so easily into complicity. This isn't merely an AI ethics kerfuffle; it's a testament to a deeply entrenched pattern. Recall the CIA's 1954 coup in Guatemala, orchestrated to protect the United Fruit Company's profits, or the 2002 attempted coup against
Hugo Chávez in Venezuela, quietly backed by Washington. These interventions, cloaked in rhetoric about 'democracy promotion,' consistently serve economic and strategic interests, often leveraging private sector tools. Here, the 'private sector tool' in question, Claude, was made accessible through Anthropic's partnership with Palantir Technologies, a company notorious for its deep, lucrative ties
to intelligence agencies and military contractors, boasting over $1.5 billion in federal contracts by 2021. The financial incentives are glaring: Anthropic gains prestige and integrates deeper into the immensely profitable national security apparatus, while Palantir further entrenches itself as the indispensable digital backbone of US power projection. The 'ethical AI' narrative, in this light,