Anthropic Rejects Pentagon's AI Safeguard Removal Demand

Anthropic will not remove safeguards from its AI systems, despite a dispute with the Pentagon. The Pentagon requested the removal of these safeguards, threatening to label Anthropic a "supply chain risk" and end a $200 million contract. Anthropic's refusal is based on preventing its AI from being used for autonomous weapons targeting or mass surveillance in the United States. Pentagon spokesperson Sean Parnell stated the department does not intend to use AI for these purposes and asked Anthropic to allow its model for "all lawful purposes." The company had until Friday evening to decide. Anthropic CEO Dario Amodei said the company opposes using its AI for mass domestic surveillance or fully autonomous weapons, citing that "frontier AI systems are simply not reliable enough." The company believes AI systems are too unpredictable for "life-or-death targeting" and could lead to unintended consequences. They also argue that AI for mass surveillance could violate the spirit of constitutional protections. The Pentagon, through Undersecretary of Defense Emil Michael, accused Amodei of lying and having a "God-complex," stating the department will "ALWAYS adhere to the law." Anthropic maintains its position, stating they "cannot in good conscience accede to their request." This standoff highlights a growing tension over the ethical use of advanced AI in defense.