This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A series of robberies and deaths in New York City involved perpetrators drugging victims at gay bars to incapacitate them. Once the victims were unconscious, the perpetrators used the victims' own faces to unlock their smartphones via facial recognition, subsequently accessing financial apps to steal funds.
Gay men in New York City were drugged by robbers who accessed their phones using facial recognition while they were unconscious to transfer funds out of their bank accounts.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Vulnerabilities that can be exploited in AI systems, software development toolchains, and hardware, resulting in unauthorized access, data and privacy breaches, or system manipulation causing unsafe outputs or behavior.
Also classified as
The perpetrators used the biometric bypass to gain unauthorized access to financial apps and steal tens of thousands of dollars from the victims.
The criminals exploited a security vulnerability in the smartphone's facial recognition system, which allowed unauthorized access to the device while the victim was unconscious.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The harm was caused by human criminals who intentionally drugged victims and exploited the deployed facial recognition system to rob them.
Added categories peak at 2: Minor (Psychological Harm, Indirect)
A series of localized robberies in New York City involved criminals drugging victims to bypass smartphone facial recognition security and steal funds. While resulting in tragic fatalities and targeted community victimization, the incident is a domestic criminal matter with minor national security implications, highlighting consumer biometric vulnerabilities rather than state-level threats.
The smartphone's facial recognition system operated as a deterministic tool, performing the specific function of unlocking the device when presented with the victim's face, without independent planning or context retention.
High Risk: The incident involves facial recognition technology, which has significant implications for personal safety, security, and fundamental rights, and is explicitly categorized under high-risk applications.
The primary function of the AI system exploited in this incident was face recognition, used to verify the identity of the user to unlock the smartphone.
The reports from reputable news organizations provide consistent, detailed accounts of the crimes, the specific mechanism of exploiting facial recognition, and the financial and physical harms suffered by the victims.
The incident could be viewed purely as a physical crime (drugging and robbery) where facial recognition was incidental, but the systematic exploitation of the biometric unlock is central to how the financial theft was executed.
Technology
Flawed Biometric Authentication
Facial recognition allowed access to the device while user was unconscious.
Lack of Multi-Factor Security
Financial apps opened immediately without secondary PIN or password checks.
Data Inputs
Unconscious Biometric Input
The system accepted facial input data without verifying active consciousness.
Human Factors
Physical Incapacitation
Victims were drugged, preventing them from stopping unauthorized phone access.
Process and Methods
Insecure Password Reset Process
Attackers changed iCloud passwords using only the unlocked phone device.
Deficient Fraud Prevention
Apps allowed immediate draining of accounts without flagging anomalous activity.
Regulatory Environment
Inadequate Law Enforcement
Police initially dismissed the tech-facilitated robberies as simple overdoses.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
