A Black Uber Eats delivery driver was dismissed after facial recognition technology repeatedly failed to verify his identity, leading to allegations of racial bias in the AI system.
Pa Edrissa Manjang worked for Uber Eats between November 2019 and April 2021 as a delivery driver while employed full-time as a financial assistant. Initially, facial verification checks were infrequent, but they became more frequent over time. Manjang was eventually dismissed by email in April 2021, with Uber claiming there were 'continued mismatches' between the pictures he took to register for shifts and his Uber work profile photo. When Manjang requested human review of the photos, Uber told him that after 'careful consideration' his account was being deactivated. Manjang's legal claim, supported by the App Drivers and Couriers Union (ADCU), alleges he was dismissed due to racially biased facial recognition technology and was subjected to heightened and excessive facial recognition verification checks that amounted to racial harassment. On Wednesday, an east London employment tribunal refused Uber's application to strike out the discrimination claim. The case also mentions another former Uber driver, Abiodun Ogunyemi, who filed a similar employment tribunal case in October 2021 after Uber's facial recognition software failed to recognize him. AI experts note that facial recognition systems are generally less accurate at identifying people from ethnic minority backgrounds due to training data that skews toward white men. Uber maintains that automated facial verification was not the reason for Manjang's dismissal and that their system includes human review.
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
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed