Behind NAZA: How AI is used to assess, reduce collateral damage in IDF – interview

The documentary "NAZA," which premiered at the Venice Film Festival, examines the Israeli military's use of artificial intelligence (AI) to assess and reduce collateral damage in Gaza. While the film's synopsis suggests Israel's systems facilitate "calculated mass killing," the IDF states these assessments are intended to minimize civilian harm by estimating incidental damage, comparing it with military advantage, and considering measures to reduce harm. Cyber intelligence expert Matías Rott explained that AI in this context is a decision-support function, integrating intelligence and modeling surroundings to help minimize collateral damage. He stated, "When you are working with this AI, for sure what you are willing or intending is to minimize collateral damages."

The film raises questions about the accuracy of these calculations, the permissiveness of civilian casualty thresholds, and the extent of human oversight. The IDF rejects the film's allegations, emphasizing that AI tools like Lavender and Habsora are not autonomous targeting machines and that the final decision always rests with a human. Rott noted that the key dilemma is whether the human acts "as a rubber stamp" or genuinely processes information. The IDF has also challenged the film's reliance on anonymous sources and is investigating potential leaks of classified material.