Uber used an AI-powered tool called Greyball to systematically deceive law enforcement officials by showing them fake app interfaces with ghost cars to evade regulatory enforcement in cities where its service was banned or restricted.
From 2014 onwards, Uber deployed a sophisticated deception program called Greyball as part of its VTOS (violation of terms of service) system to evade law enforcement in cities where its ride-hailing service was banned or restricted. The system used geolocation data, credit card information, social media accounts, and device identification to identify law enforcement officials conducting sting operations. When identified officials tried to hail rides, they were shown fake versions of the Uber app populated with ghost cars that would never arrive, or drivers would be instructed to cancel rides. The program was used in cities including Portland, Boston, Philadelphia, Las Vegas, and internationally in France, Australia, China, South Korea, and Italy. In Portland in late 2014, code enforcement inspector Erich England was successfully deceived during a sting operation, unable to catch any Uber drivers. The program was approved by Uber's legal team and known by at least 50 people within the company. Legal experts suggested the program could violate the Computer Fraud and Abuse Act or constitute intentional obstruction of justice. The revelation came amid broader scandals at Uber including sexual harassment allegations and executive resignations.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
AI system
Due to a decision or action made by an AI system
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed