AI algorithms used by companies like Uber, Lyft, and Amazon to set personalized wages for gig workers based on extensive data collection, resulting in differential pay for the same work and creating what researchers term 'algorithmic wage discrimination.'
According to research by UC Hastings law professor Veena Dubal published in January 2023, companies like Uber, Lyft, and Amazon use AI algorithms to implement 'algorithmic wage discrimination' against gig workers. The study, based on six years of interviews with hundreds of drivers, found that these companies collect massive datasets on contract workers including when they work, duration, and past payment acceptance patterns. The AI systems use this data to calculate personalized wage rates designed to incentivize specific behaviors, offering different workers different pay for identical work. For example, algorithms might offer lower rates to drivers predicted to accept them, or systematically reduce rates for drivers approaching daily earning targets to extend working hours. One interviewed driver, Domingo, reported experiencing 45 minutes of dead time when one ride short of a $100 bonus, describing it as feeling 'punished by some unknown spiteful God.' Workers described the system as 'gamblification of work' and compared it to gambling where 'the house always wins.' The California Labor Commission cited this practice among reasons for suing Uber and Lyft in 2020, claiming $1.3 billion in owed payments. Uber denied the allegations, stating they do not tailor individual fares based on driver willingness to accept low pay.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Unequal treatment of individuals or groups by AI, often based on race, gender, or other sensitive characteristics, resulting in unfair outcomes and unfair representation of those groups.
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