The authoritarian AI crisis has arrived

AI safety researchers have long worried that a government would seek to use AI for domestic surveillance and autonomous killing. The Pentagon’s fight with Anthropic threatens to make it a reality