"In Iran, AI has potentially already been involved in identifying exponentially more targets than in previous wars, said Utrecht University’s Dorsey. Those targets could have existed beforehand — or they could have been generated quickly by AI systems, creating a serious concern about how carefully these have been vetted as required by law, she said.“How do you lift the veil on a system making 37mn computations per second? How on earth would you even be able to even trace that back in any way?” Dorsey said. “Are you going to meaningfully exercise context-appropriate human control and judgment over decisions that are generated by these systems?”"https://www.ft.com/content/fedb262e-e6db-40bc-a4d0-080812f0f82b#Iran #USA #KillChain #AI #Palantir #Anthropic
