Mass Facial Recognition Program in Gaza Reportedly Used by Israeli Forces to Identify Palestinians

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.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.3 Lack of capability or robustness

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.

Additional risk subdomains

• 4.1 Disinformation, surveillance, and influence at scale: The military deployed the AI system to conduct mass surveillance, collecting and cataloging the faces of Palestinians without their knowledge or consent.
• 2.1 Compromise of privacy by obtaining, leaking or correctly inferring sensitive information: The program built a database of personal biometric data (facial scans) of Palestinians without their consent.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

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.

EU AI Act risk tier

• Risk tier: 2 High Risk

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.

AI system and alleged parties

• AI system: Corsight facial recognition technology, Google Photos (Corsight, Google)
• AI purpose: Face Recognition; Target Identification
• Behaviour type: Assistant
• Alleged developer: Unknown Israeli military integrators, Google Photos, Corsight
• Alleged deployer: Unit 8200, Israeli military intelligence, Israeli government, Israel Defense Forces
• Alleged harmed parties: Privacy and data rights of Gaza residents, Palestinians traveling through Gaza checkpoints, Palestinians, National security and intelligence stakeholders, Mosab Abu Toha, Gazans, Epistemic integrity

Harm severity

Highest direct severity in any category: Severe. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Minor
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Substantial, indirect Substantial
• Civil rights: direct Substantial, indirect Substantial
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Severe, indirect Substantial
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Physical

Reported: The report explicitly describes physical harm to at least one individual who was detained and beaten.

Directly caused: N/A

Indirectly caused: The poet [PERSON_001] was detained for two days, beaten, and interrogated by Israeli officers after the AI system mistakenly flagged him as a wanted person.

Inferred additional harm: It is highly likely that other individuals among the 'hundreds' flagged by the system also experienced physical abuse, detention, or interrogation, potentially affecting dozens of people.

Differential treatment

Reported: Yes, the report describes systematic differential treatment of Palestinians in Gaza.

Directly caused: The AI facial recognition database was specifically built to catalog and track Palestinians without their knowledge or consent.

Indirectly caused: Palestinians fleeing conflict zones were subjected to biometric screening and potential detention at checkpoints, a measure not applied to other groups.

Inferred additional harm: The deployment of experimental biometric surveillance targeting an entire occupied population likely reinforces systemic discrimination and unequal treatment on a massive scale.

Civil rights

Reported: Yes, the report explicitly describes violations of human and civil rights.

Directly caused: The AI system was used to conduct mass surveillance and catalog the faces of Palestinians without their knowledge or consent, violating their right to privacy.

Indirectly caused: The poet and hundreds of others faced arbitrary detention, interrogation, and physical abuse due to false positive matches by the AI system, violating their rights to liberty and freedom from torture.

Inferred additional harm: Thousands of Palestinians likely suffered violations of their rights to privacy, freedom of movement, and due process due to the unchecked deployment of this automated targeting and surveillance system.

Privacy

Reported: Yes, the report explicitly describes massive privacy violations.

Directly caused: The AI program collected, cataloged, and built a database of the facial biometrics of Palestinians without their knowledge or consent.

Indirectly caused: The military used Google Photos and Corsight to search and identify individuals from crowds and drone footage, compromising their expectation of privacy in public spaces.

Inferred additional harm: The biometric data of hundreds of thousands of Palestinians has likely been permanently cataloged in military databases, representing a persistent, long-term violation of privacy.

Psychological

Reported: The report describes psychological distress and trauma resulting from wrongful detention and interrogation.

Directly caused: N/A

Indirectly caused: The poet experienced severe confusion and distress, stating he had 'no idea what was happening' when blindfolded and led away for interrogation based on the AI's identification.

Inferred additional harm: Widespread anxiety, fear, and trauma are likely experienced by the hundreds of Palestinians subjected to sudden, automated checkpoint detentions and the broader population living under constant biometric surveillance.

People affected

• Occurrences reported: 1
• People reportedly harmed: 200
• People reportedly exposed: 10000

Potential causes

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.

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.

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.

Information quality

• Classification confidence: High
• Reason for confidence: 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 [PERSON_001].
• Ambiguities identified: The exact number of false positives and the total number of people detained solely due to AI errors remain unspecified.
• Alternative interpretations: 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.

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.

• Overall national security impact: Severe
• Response level: Severe
• Scope: Multiple nations
• Primary target: Palestine
• Alleged perpetrator: Israel

Threat characteristics

• Imminence: Long-term. Ongoing military capability and surveillance program rather than an isolated, immediate crisis.
• Autonomy: Human-supervised. AI autonomously scans and flags targets, but human soldiers execute detentions and interrogations.
• Novelty: Evolved capability. Significant advancement in scaling facial recognition and consumer tech for military targeting in active conflict zones.

Impact by dimension

• Physical security: Substantial. AI-enabled target identification led directly to wrongful military detention, interrogation, and physical abuse of civilians at checkpoints.
• Information security: Substantial. Mass biometric surveillance program deployed by military intelligence, integrating commercial and private AI tools to scan and catalog population without consent.
• Sovereignty: Substantial. Deployment of experimental foreign military AI surveillance systems within occupied territory, severely disrupting civilian freedom of movement.
• Economic security: Minor. Misuse of commercial consumer AI technology (Google Photos) integrated into military intelligence operations, raising tech dual-use risks.
• Societal stability: Severe. Mass automated surveillance and systemic human rights violations, including arbitrary detention and physical abuse of civilians due to algorithmic errors.