This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Uber drivers in the UK and Portugal launched legal challenges against the company for 'robo-firing' workers based on automated fraud detection systems. The Amsterdam Court of Appeals ruled that Uber's automated processes for deactivating drivers and managing pay/work allocation constitute automated decision-making under GDPR, requiring transparency and meaningful human intervention. The court found that Uber's human oversight was largely symbolic and ordered the company to disclose the logic behind its algorithmic management and fraud probability scores.
Uber was alleged in a lawsuit to have wrongfully accused its drivers in the UK and Portugal of fraudulent activity through automated systems, which resulted in their dismissal without a right to appeal.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Delegating by humans of key decisions to AI systems, or AI systems that make decisions that diminish human control and autonomy, potentially leading to humans feeling disempowered, losing the ability to shape a fulfilling life trajectory, or becoming cognitively enfeebled.
Also classified as
The use of algorithmic management and robo-firing reduced the quality of employment for gig workers, exacerbating power imbalances.
The automated fraud detection systems falsely flagged innocent drivers for fraudulent activity, demonstrating a lack of robustness in critical employment decisions.
The ride-hailing platforms delegated termination and management decisions entirely to automated algorithms, removing meaningful human review and leaving drivers with no recourse.
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
The algorithmic management and automated deactivations were executed by deployed AI systems designed by Uber and Bolt to manage their workforce.
Added categories peak at 2: Minor (Psychological Harm, Indirect)
Gig economy drivers successfully challenged automated termination and management algorithms under GDPR. The Amsterdam Court of Appeals ruled that Uber's human oversight of its algorithmic deactivations was merely symbolic. While representing a significant precedent for labor rights and algorithmic transparency, the national security impact remains negligible to minor.
The reports describe the AI systems making independent decisions to deactivate driver accounts and allocate work without meaningful human intervention. The court found that human review was merely a 'symbolic act', meaning the system operated autonomously in practice.
High Risk: The AI system is used for employment and worker management, specifically for deactivating accounts and managing pay. The definitions state that 'Employment and worker management systems, such as AI used in recruitment' are classified as High Risk.
The AI system's primary function is to monitor driver behavior, evaluate performance, calculate fraud probability scores, and administer the workforce by allocating rides and deactivating accounts.
The reports provide detailed, consistent accounts of the legal proceedings, the specific rulings of the Amsterdam Court of Appeals, and the arguments from both the drivers' representatives and the platforms. The facts regarding the lack of human intervention and the types of data withheld are well-documented across multiple sources.
Uber argued that its processes had meaningful human involvement and did not constitute solely automated decision-making, though the court rejected this defense.
Technology
Opaque Algorithmic Deactivation
Algorithms automatically terminated driver accounts based on suspected fraud.
Flawed Fraud Detection System
Algorithmic systems flagged normal driver behavior as fraudulent activity.
Black Box Dynamic Pricing
Complex algorithms calculated prices and pay with zero transparency.
Data Inputs
Inaccessible Driver Profiles
Platforms withheld passenger ratings and electronic performance tags.
Hidden Fraud Probability Scores
Drivers were denied access to algorithmic fraud scores used to manage them.
Incomplete Subject Data
Platforms provided incomplete data, omitting GPS and trip-level ratings.
Human Factors
Symbolic Human Review
Human intervention in deactivations was a symbolic act with no real check.
Lack of Driver Recourse
Drivers had no way to explain their side or appeal automated decisions.
Unqualified Review Staff
Staff in Krakow lacked qualifications to properly review automated firings.
Process and Methods
Stonewalling Data Requests
Platforms denied data access requests, citing trade secrets and security.
No Formal Appeals Process
Drivers were summarily dismissed with no mechanism to challenge errors.
Obfuscated Pay Structures
Fragmented pay structures made it impossible to verify actual earnings.
Management
Prioritizing Automation Over Fairness
Platforms built automated HR systems to save money and avoid duties.
Power Asymmetry Exploitation
Management relied on information asymmetry to control the workforce.
Deflecting Cybersecurity Failings
Management blamed platform security issues on drivers via false fraud claims.
Regulatory Environment
Weak Regulatory Oversight
Regulators like TfL failed to review major algorithmic pricing changes.
Threat of Stripped Protections
Proposed UK data bills threaten to remove protections against robo-firings.
Inadequate Platform Directives
Lack of harmonized platform work regulations allowed worker exploitation.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
