An Uber Eats bike

Uber Eats sued for ‘racist’ facial recognition software

Uber Eats is being sued by one of its employees over its new facial recognition app. Pa Manjang, who described the app as ‘racist’, is of African black origin, and moved to the UK from Gambia in 2011.

Uber Eats requires couriers to take photos of their face every day, verifying their identity and that the delivery is being carried out by them specifically. Once uploaded, the photo is then compared to the original ID of the courier.

Manjang claimed the automated software rejected his uploaded photo on numerous occasions as it would indicate his selfies were of a different person. Manjong added that he was ‘asked to take photos of [himself] multiple times in a day’ because of the facial recognition system not being able to validate his identity.

The Face API facial recognition software is a Microsoft Cognitive Services product. Uber Eats uses the face ID verification system to prevent fraud and as an added security feature of its service. The system works by comparing two images of a person’s face, seeking to evaluate their degree of similarity.

Filters generated by a computer convert the images into numerical expressions which are then compared for pattern-identification. The software is meant to use machine learning processes that study the data that feeds into its system, and over time configure to recognise patterns and expressions accurately and more quickly, which in this case is the recognition of people’s faces.

Uber Eats, like many other delivery companies, uses Face API to ensure the courier making the delivery is the individual registered to make the delivery. However, as seen in Manjang’s case, the software can have inaccuracies.

In April 2021, Uber Eats permanently suspended Manjang, claiming he had shared his account as the facial recognition app had failed to correctly register his picture and match to his ID photo.

The case brought to the Employment Tribunal by Manjang was met with fierce rejection and denial after Uber Eats tried to have the case thrown out. Uber London Ltd and Uber Portier BV also attempted to have the case dismissed by the judge, refuting claims that they had racially harassed and discriminated against Manjang.

However, an East London judge has held onto the case and the lawsuit continues.

The suspended London driver said the recognition app is racially-biased as a ‘plethora of research exists’ to suggest false positive and false negative results. He claims the software disadvantages people of ethnic minority.

Whether this case prompts delivery businesses to stop depending on AI software for delivery courier verification or push Microsoft to develop its technology to reduce racial bias, the moral importance in Munjang’s story is the employer’s role and responsibility to ensure the wellbeing of their employee.

Employers have a duty of care to their workers and must do all they can to protect them from workplace discrimination. Race is a protected characteristic under the Equality Act and should Uber Eats as an employer be causing racial discrimination due to its service operations and processes to its employees, great legal implications potentially exist.