What Israel's use of AI in Gaza revealed about their civilian harm thresholds

Bloomberg defense tech reporter Katrina Manson, author of a new book on Project Maven, discusses what the IDF's use of AI in Gaza reveals about the gap between US and Israeli standards for civilian harm. Manson draws on investigative reporting from 972 Magazine, as well as her own conversations with US military officials who analyzed the IDF's approach.

That reporting found very little time between AI-surfaced intelligence and actual strikes. US officials told Manson that Israel had not broken the laws of armed conflict, but was "leaning in towards autonomous processes in a way that the US had never leaned anywhere near as close."

What this reveals is a stark difference in how two allied militaries weigh civilian harm. Manson says the IDF was prepared to accept higher casualty numbers per mid-ranking target than US military tradition allows. That gap is a policy question as much as a technology one; as Manson and Drew Kukor, the former chief of Project Maven, both acknowledge, AI and policy are "absolutely interlinked".