Pentagon bets speed beats safety flaws
the us navy adopted a new 'ai-first' strategy declaring that slow adoption poses a greater risk than 'imperfect alignment'. this signals a wartime shift where speed of deployment trumps theoretical ai safety concerns in military operations.
the department of the navy approved a 'strategy to weaponize data and artificial intelligence', introducing a five-stage 'bits2effects cycle' designed to minimize the time between data capture and military action. acting secretary hung cao signed the directive to build an 'ai-first' force, aiming to double data and ai engineers by fy2029 and run llms directly on vessels. notably, the strategy adopts a 'wartime approach', asserting that the risks of moving slowly outweigh those of 'imperfect alignment' in ai systems. this prioritizes rapid operational capability and adaptation over strict safety guarantees, marking a significant shift in military acquisition and risk tolerance.