Tel Aviv: Amidst the conflict with Hamas in Gaza, the Israeli military has employed artificial intelligence to aid in pinpointing bombing targets in the war-ravaged areas, as reported by CNN citing an investigation by +972 Magazine and local insights from six Israeli intelligence officials involved in the purported program.
As per the Israeli officials, disclosed in a comprehensive probe by the online publication co-managed by Palestinians and Israelis, the AI-driven tool dubbed “Lavender” exhibited a 10 percent margin of error.
When prompted about +972 Magazine’s exposé, the Israel Defence Forces (IDF) neither refuted the tool’s existence nor confirmed AI’s application in identifying suspected terrorists. However, in a detailed statement, it stressed that “information systems serve merely as aids for analysts in the target identification process,” and Israel endeavors to “minimize harm to civilians to the fullest extent possible within the operational circumstances prevailing during the strike.”
CNN also cited the Israel Defence Forces (IDF) stating that “analysts are required to conduct independent assessments, ensuring that the identified targets align with the relevant definitions as per international law and additional restrictions outlined in the IDF directives.”
Nevertheless, one official disclosed to +972 that human personnel often functioned merely as a “rubber stamp” for the machine’s decisions, typically allocating only about 20 seconds to each target – confirming their gender – before sanctioning a strike.
This investigation arises amid heightened international scrutiny of Israel’s military operations following targeted airstrikes that claimed the lives of several foreign aid workers distributing provisions in the Palestinian enclave. According to the Gaza Ministry of Health, Israel’s blockade of Gaza has resulted in over 32,916 casualties and triggered a deepening humanitarian crisis, with nearly three-quarters of the population in northern Gaza enduring catastrophic levels of hunger, as per a United Nations-supported report.
The investigation’s author, Yuval Abraham, previously disclosed to CNN in January the extent to which the Israeli military has been “heavily reliant on artificial intelligence to generate targets for such operations with minimal human oversight.”
Contrary to assertions, the Israeli military “does not employ an artificial intelligence system tasked with identifying terrorist operatives or forecasting whether an individual is a terrorist,” stated the IDF on Wednesday. However, its analysts utilize a “database designed to cross-reference intelligence sources to generate current information layers on the military personnel of terrorist entities.”
Subsequently, human officers are tasked with confirming “that the identified targets comply with the relevant definitions as per international law and additional constraints outlined in the IDF directives,” as described in the IDF statement, a process similarly outlined by +972.