Documentary 'Nazha' says Israeli AI systems picked targets in Gaza war

Artificial intelligence, once meant to make everyday life easier in fields like education, healthcare, business and research, is now being used to make life-and-death decisions on the battlefield. Wars are no longer fought only with guns, missiles, aircraft and soldiers; computer algorithms and AI are now telling armies whom to kill and whom to spare.

A documentary called 'Nazha', on the Israel-Gaza war, premiered recently at the Venice International Film Festival. Its release has triggered widespread anger globally over how technology could make such ruthless decisions about human life and death. The documentary shows how the Israeli military used AI technology to locate and target people during the war.

The makers of 'Nazha' cite 24 separate sources from the Israeli Defence Forces (IDF) and intelligence agencies. The sources' identities have been kept secret for security reasons, but their detailed accounts reveal two powerful and secret Israeli AI systems, which the documentary claims were used extensively during the Gaza war.

The first system, Lavender, analysed phone data, messages and suspicious activity of people in Gaza to automatically generate a list of "terrorist" and military targets. The second, Where's Daddy, kept round-the-clock watch on suspects flagged by Lavender. It tracked when a person reached home to be with their family, and sent real-time alerts to Israeli military operators.

A former Israeli military official interviewed in the documentary said that in the early months of the war, the Lavender system had drawn up a target list of about 37,000 potential suspects. Before October 7, the AI was trained on data about Hamas operatives' phones, conversations and daily movements. It then scanned phone data across the whole of Gaza and generated the list itself.

The documentary claims that Israeli military operators were supposed to carry out a human check before giving final approval to a target chosen by the AI system. But in practice, each operator was given only 20 seconds to review a target. In such a short time it was impossible to tell whether the AI's decision was right or wrong, so operators simply rubber-stamped the AI's call. The documentary alleges that this led to the deaths of many innocent people, including women and young children, because of the AI's faulty analysis.

Following the documentary, the Israeli Defence Forces (IDF) have completely rejected all its claims. A senior military official said the information shown in the documentary was exaggerated and paints the military in a false light. However, the official did not entirely deny the use of AI and advanced computer technology in modern warfare. The Israeli military says that before giving final approval for any air strike or operation, military officials carefully examine the actual situation on the ground, the presence of civilians nearby, the risk to life and the military objective.