IDF Denies AI Makes Strike Decisions, Cites Human Oversight
The Israel Defense Forces (IDF) has refuted claims that artificial intelligence systems are making targeting decisions, asserting that such choices are made by highly qualified military personnel. In response to accusations in the documentary "NAZA," Israeli military officials told the Jerusalem Post that air force officers and military legal advisors are involved in the strike authorization process. While AI may be used as an analytical tool, it does not replace human judgment, they stated.
The IDF also criticized the documentary's creators for not sharing the film or specific accusations prior to its release, which prevented the military from providing detailed responses. The army acknowledged civilian casualties in Gaza but maintained that measures are taken to minimize them. They accused Hamas of using civilians as human shields and criticized the documentary for omitting IDF efforts like organized evacuations and strikes canceled due to civilian presence.
According to the Jerusalem Post, the air force has, in some instances, diverted missiles already launched when civilians unexpectedly appeared near a target. Military officials admitted that errors can occur in complex, urgent operations, citing the June 2024 hostage rescue mission where the situation rapidly deteriorated. The IDF also disputes casualty ratios suggesting 100 to 500 civilians killed per Hamas operative, deeming them inaccurate, though they acknowledged some strikes resulted in heavy tolls due to errors or incomplete intelligence.
Disagreements among officers regarding the appropriateness of certain strikes, particularly after March 2025, were acknowledged. However, officials characterized these as internal professional debates, not evidence of AI-driven targeting.
