This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Israeli military intelligence deployed a facial recognition program in Gaza using Corsight technology and Google Photos to identify and detain individuals. The system, which operated without the knowledge or consent of the Palestinian population, produced false positives that led to the arbitrary detention and interrogation of civilians, including a Palestinian poet. The program was used to create 'hit lists' of suspected militants, but its reliance on grainy footage and automated matching resulted in significant errors.
A previously undisclosed facial recognition initiative operated by Israeli military intelligence units was reportedly deployed across Gaza after the October 7, 2023 attacks. According to multiple intelligence officers, the program uses Corsight technology alongside Google Photos to identify individuals from checkpoints, crowds, and drone footage. The system has allegedly produced misidentifications, including the widely reported detention of Palestinian poet Mosab Abu Toha on November 19, 2023.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
Also classified as
The program built a database of personal biometric data (facial scans) of Palestinians without their consent.
The military deployed the AI system to conduct mass surveillance, collecting and cataloging the faces of Palestinians without their knowledge or consent.
The AI system failed to perform reliably under varying conditions (such as grainy footage or obscured faces), resulting in false positives that wrongly flagged innocent civilians as terror suspects.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The primary harm was caused by the AI system's false positive identification of an innocent civilian, which was an unintentional error during its post-deployment operation.
Added categories peak at 3: Substantial (Psychological Harm, Inferred)
The deployment of AI-driven facial recognition by the Israeli military in Gaza represents a significant escalation in automated mass surveillance. The system's high error rate led to the arbitrary detention and physical abuse of innocent civilians, demonstrating severe human rights implications and the risks of deploying unvalidated AI targeting tools in conflict zones.
The AI systems (Corsight and Google Photos) functioned as assistants to the military, scanning faces and flagging suspects to augment human intelligence efforts, while still requiring soldiers to perform the physical detentions and interrogations.
Risk Level 2. High Risk: The system is a law enforcement and military application of facial recognition, which is explicitly categorized as a High Risk application under the provided definitions.
The primary function of the AI systems was to perform facial recognition on crowds and drone footage to identify individuals, with the secondary goal of identifying military targets.
The reports from reputable sources (The New York Times, The Verge) provide consistent, detailed accounts of the deployment, the specific technologies used (Corsight and Google Photos), and the direct consequences on individuals like the poet [name withheld].
The military might argue the system is a standard intelligence tool rather than an experimental AI deployment, but the reports strongly emphasize its novel, automated, and error-prone nature.
Technology
Algorithmic Inaccuracy
Corsight struggled with grainy footage, causing false matches.
Unsuited Consumer Software
Google Photos was used as an ad-hoc military tool despite being consumer-grade.
Data Inputs
Poor Quality Imagery
Grainy drone and security footage reduced the reliability of facial matches.
Unverified Informant Data
Target lists relied on unverified names provided by Palestinian prisoners.
Human Factors
Automation Bias
Soldiers rarely questioned algorithm matches, leading to wrongful detentions.
Lack of Informed Consent
Targeted individuals had no knowledge that their faces were cataloged.
Process and Methods
Broad Targeting Guidelines
Vague rules for stopping individuals led to wide margins of error.
Inadequate Verification Process
No robust process to verify AI matches before interrogation and detention.
Management
Prioritizing Speed Over Accuracy
Unit 8200 prioritized rapid creation of hit lists over system reliability.
Ad-hoc Tool Deployment
Management permitted unvetted use of Google Photos in military workflows.
Regulatory Environment
Lack of Legal Oversight
Mass surveillance program operated without regulatory scrutiny or legal limits.
Absence of Use Restrictions
No regulations prevented military deployment of commercial facial recognition.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
