“Well, the drones would come down and pick off civilians, children…This is not, you know, an occasional thing. This was day after day after day.” – Nizam Mamode, a volunteer surgeon at Nasser Hospital in Central Gaza, in an interview reported by NPR.
This is what it’s like to live in Gaza. In September 2025, the United Nations Commission of Inquiry on the Occupied Palestinian Territory, an independent investigative body established by the United Nations Human Rights Council following a vote on 27 May 2021 to establish a fact-finding mission to investigate war crimes in occupied Palestine, concluded that Israel is committing genocide against Palestinians in the Gaza Strip–a finding the Israeli government has rejected, and one that remains distinct from any final ICJ ruling. The independent commission found that Israel has violated 4 out of 5 acts of genocide under the 1948 Genocide Convention. Additionally, the International Association of Genocide Scholars, the world’s largest professional organization of genocide scholars, including numerous Holocaust experts, concluded that “Israel’s policies and actions in Gaza meet the legal definition of genocide,” with 86% of voting members in agreement in a formal resolution passed in August 2025. As Israel’s current genocide against Palestinians continues to unfold, new technologies, including AI (artificial intelligence) warfare, are being tested on Palestinians as a tool to further escalate the mass killing. Semi-autonomous AI targeting systems are driving kill decisions with minimal human oversight, and private companies such as Palantir are profiting from it and continue to face zero accountability. This raises the question: What can we expect for the future of warfare?
In Gaza, AI algorithms are using surveillance and location data in phones to target civilians. “Lavender” is an AI tool designed to give Palestinians scores from 1 to 100 for their likelihood of being suspected militants. At the start of the genocide, the system flagged as many as 37,000 Palestinians as “Hamas militants.” In addition, an investigation by +972 and Local Call has shown that Lavender achieved only 90% accuracy in target selection. That means that of all of the flagged Palestinians, 10% weren’t militants at all. According to testimonies of six anonymous Israeli intelligence officers, they admitted that they would only spend 20 seconds checking each target flagged by Lavender, only to confirm it was male. This demonstrates that there was little to no true human oversight on these lethal decisions. The IDF argues that tools like Lavender are simply supplemental tools, insisting that each target is reviewed and analyzed for legitimacy by a human. This contradicts the testimonies from numerous intelligence officers mentioned earlier.
On top of Lavender, Israel additionally uses a system called “Where’s Daddy,” which tracks flagged Palestinians and waits for them to enter their homes before they are bombed, with just 20 seconds of human review. In the investigation, an Israeli intelligence officer said, “The IDF bombed them [Palestinian targets] in homes without hesitation, as a first option. It’s much easier to bomb a family’s home. The system is built to look for them in these situations.” According to The Office of the High Commissioner for Human Rights, the UN’s human rights office, a large proportion of those killed in Gaza were in residential buildings. When Israel utilizes these various computer systems, they don’t just kill the flagged Palestinians but also innocent family members. The machine handles all the finding, tracking, and targeting; all that’s left for humans is to click approve. Behind the machine, a company is collecting a check.
Lavender and Where’s Daddy are not the only AI tools being used. The broader reality is that private companies are openly selling battlefield AI to the IDF and profiting from supporting a genocide. Palantir is a private American firm with partnerships with Israel’s Ministry of Defense to provide software to support military operations and battlefield decision-making. The company has openly supported Israel’s military actions. Although there is no direct link between Palantir and the “Lavender” or “Where’s Daddy” systems, the company is confirmed to be providing battlefield AI software to the Israel Defense Forces (IDF). Semi-autonomous targeting systems like these make it easier for militaries and private companies to hide behind AI. Current international law requires a real human being, not a machine, to retain meaningful control over the decision to take a life. Article 36 of the Geneva Convention requires legal review of all new weapons, but it was written 47-48 years ago, before AI existed. It establishes no clear threshold for what constitutes meaningful human oversight. While there are said to be humans in the loop, a 20-second gender check is not a meaningful control under any reasonable standards. Yet nothing in existing international law prohibits it. This exposes a gap in international law; decisions made by Lavender are simply stamped for approval by humans, yet no existing international law defines this as a violation.
Gaza has long been used as a testing ground, and Palestinians are the test subjects. This is nothing new, as journalist Antony Loewenstein argues that Gaza and the West Bank have been used as a laboratory for surveillance technology and weapons for decades. Israeli technology was later labeled as “battle-tested” and sold to over 130 countries, including use at the US-Mexico border. The use of AI targeting systems will not be confined to Gaza; these technologies won’t stay in the laboratory. As entire Palestinian families are being wiped out by AI, international humanitarian law needs to set rules now, before AI warfare is adopted worldwide.
