‘Naza’: In new documentary, Israeli military whistleblowers expose how civilian mass killings were part of IDF’s Gaza calculations
A widely circulated still from the documentary “Naza”.
New Delhi: A documentary unveiled at the Venice film festival on Thursday alleges that Israeli military targeting decisions routinely incorporated large civilian losses during the Gaza war. Directed by Israeli filmmakers Yuval Abraham and Rachel Szor, “Naza” presents testimony challenging official accounts of how attacks were planned and approved.
The Guardian, which produced the film, says it features 24 Israeli military insiders. Its title is an acronym used by Israeli intelligence for the number of civilians expected to die in an airstrike. The Hebrew term is “nezek agav” (plural is “nezek agabi”), which means “damage/harm incidental/collateral”.
Interviewees describe phone surveillance supporting strikes on homes and killings near undisclosed exclusion zones. The Israeli military had not immediately responded to Reuters’ questions about the documentary; Israel says it takes extensive measures to avoid civilian casualties.
Filmed at night on Tel Aviv rooftops, “Naza” builds on investigations published by +972 Magazine, the Hebrew-language outlet Local Call and the Guardian between 2023 and 2025. It is the only documentary competing for the Golden Lion at this year’s Venice festival.
Abraham and Szor previously co-directed “No Other Land”, the account of Palestinian displacement in the occupied West Bank that won the documentary Oscar in 2025. Their latest work shifts attention towards military personnel’s accounts of the conduct of the Gaza campaign.
How targeting systems reportedly worked
An investigation published by +972 Magazine and Local Call on April 3, 2024, described “Lavender” , an AI-based tool that flagged suspected Hamas and Palestinian Islamic Jihad operatives. Six Israeli intelligence officers said it generated as many as 37,000 potential targets during the war’s early stages.
Another tool, “Where’s Daddy?”, reportedly tracked designated individuals’ arrival at their homes. Sources alleged that suspected junior operatives were frequently attacked there at night, when their families were present, because locating them at home was easier.
One officer described spending about 20 seconds reviewing a target, largely to confirm that the person was male. Sources also alleged that commanders sometimes permitted 15 or 20 civilian deaths when attacking a suspected junior operative, although these were accounts of particular wartime practices rather than independently established rules covering every strike.
The military rejected the investigation’s characterization, saying intelligence tools assisted analysts rather than replaced their assessments. It maintained that an independent examination was required to establish whether an individual was a legitimate target.
These competing accounts place human oversight at the centre of the dispute. The question is whether personnel meaningfully examined the evidence and risks before authorizing attacks, or whether automated recommendations received little more than formal approval.
Why the technology raises wider concerns
Human Rights Watch examined the digital tools in a separate assessment published on September 10, 2024. It warned that incomplete information, unreliable approximations and excessive confidence in automated outputs could increase the danger to civilians.
Its assessment distinguished Lavender’s classification of people from another system, the Gospel, which generated potential structural targets. It also examined mobile-phone tracking used to estimate population movements and locate individuals, highlighting the limitations of such data in a territory with severely damaged communications infrastructure.
The organization’s senior surveillance researcher, Zach Campbell, warned that weaknesses in the tools’ design and application could contribute to unlawful civilian deaths. Human Rights Watch nevertheless acknowledged that it could not establish precisely when and where the systems had been deployed, or how extensively other intelligence had supplemented their outputs.
That limitation is significant when assessing individual attacks. Evidence that a military possesses a particular system does not, by itself, establish that the system generated a specific target or determined the decision to strike it.
Questions of legality and accountability
International humanitarian law requires armed forces to distinguish military targets from civilians and to take feasible precautions to reduce civilian harm. It also prohibits attacks expected to cause civilian losses excessive in relation to the concrete and direct military advantage anticipated, according to the International Committee of the Red Cross.
Applied to the allegations surrounding these systems, the unresolved questions concern identification, verification, weapon selection and the authority to approve foreseeable civilian losses. Establishing responsibility would require investigators to connect testimony with targeting records, orders, intelligence assessments and evidence from the affected locations.
“Naza” follows a string of films highlighting Israeli conduct in Gaza that have gained traction on the international festival circuit. Last year’s Venice festival also featured “The Voice of Hind Rajab”, a dramatization of the killing of a 5-year-old Palestinian girl by Israeli forces, which took the festival’s second-highest honour.
Whether “Naza” wins the Golden Lion or not, its makers say the film’s real purpose lies beyond the awards podium: to place first-hand testimony from within the Israeli military machine on the record, in the hope of shaping how governments and publics abroad judge the conduct of the war.
