Uber drivers claim they were fired after company's identification software failed to recognise their faces

  • Video report by ITV News Business and Economic Editor Joel Hills

Uber has more than 70,000 drivers in the UK. Each of their faces is unique and, thanks to modern technology, each face is now a password that Uber can use to authenticate drivers and ensure its rides are safe. In recent years, Uber has had problems with drivers faking their identities. “Real-Time ID Check” is the company’s response. The Uber app regularly asks drivers to submit a selfie before logging on. A driver can choose whether the selfie is verified by a human or facial identification software which compares it to a photograph on Uber’s database. The problem is the software isn’t always accurate.

Imran Raja says he was fired by Uber after the verification software failed to recognise his face, leaving his family without his income for three months. He says the experience had an impact on both his mental and physical health.

“Without any knowledge they revoked my licence and in the second month [thought] I’m going to go on the street with my family,”

Mr Raja told ITV News “I haven’t money to pay the rent and the car. So they broke my everything, they broke my future.” ITV News has seen copies of the messages that Uber sent Imran. On 2nd October 2020, Uber informed him it had been “unable to verify [his] identity for a second time” and had decided to "terminate the partnership”. Imran insisted an error had been made but Uber reported his dismissal to Transport for London (TFL).

On 21st October, TFL wrote to Imran and told him it no longer considered him “fit and proper” to be a taxi driver and revoked his private-hire licence "with immediate effect" A month after firing him, Uber told Imran it had finished its review of his case and “[would] not be taking any further action”.

Imran Raja says he was fired by Uber after the verification software failed to recognise his face. Credit: ITV News

Imran’s account was reactivated but he couldn't work because his licence had been cancelled. It took TFL another two months to reinstate it. Imran says Uber hasn’t apologised, neither has it offered to compensate him. Imran isn’t alone.

Abdul Kadir Ali told us that he too was dismissed by Uber after failing the Real-Time ID Check in December. Uber says his account was deactivated because he “submitted a photo of photo” in violation of the company’s policy”. Once again, TFL revoked his licence. He has now registered for universal credit. “Face ID is a good idea,” Mr Ali told ITV News. “But the system makes mistake (sic) and I lost my job and I lost my licence”.

Abdul Kadir Ali told us that he too was dismissed by Uber after failing the Real-Time ID Check in December. Credit: ITV News

The FACE API technology that Uber uses for its Real-Time ID Checks was developed by Microsoft.

Two years ago, in an interview with ITV News, the president of Microsoft said facial recognition technology has great potential but also serious limitations. “One of the challenges with the technology in its current form is it doesn’t work as well for women as it does for men, it doesn’t work as well for people of colour,” Brad Smith said.

“It is more likely to find an error, it’s more likely to find a mismatch, it is more likely to fail to identify you, it’s more likely to identify you as someone you are not.”

Many of Uber’s drivers are from minority ethnic groups.

The Uber app regularly asks drivers to submit a selfie before logging on. Credit: ITV News

The App Drivers and Couriers Union says seven members have been fired after Uber’s identification software failed to recognise them. It plans to take legal action against the company. “I think the issue is how the technology is applied and how it's governed. So the technology itself may be technically correct, but if it's not governed properly, if it's not applied properly, you're going get bad results,” says James Farrar, the union’s General Secretary. “We have a workforce here who is being discriminated against, being denied the rights. The protections are simply not there. And Uber is not in a position to govern it properly, apparently neither is TFL. I say to Microsoft, suspend the use of it.” Uber wouldn’t tell us how many times its Real-Time ID Check has failed to identify drivers since it was introduced last April. In a statement it did apologise to Imran for his experience but insists the mistake was down to “human error”. Uber said: “While no tech or process is perfect and there is always room for improvement, we believe the technology, combined with the thorough process in place to ensure a minimum of two manual human reviews prior to any decision to remove a driver, is fair and important for the safety of our platform.”

Uber has more than 70,000 drivers in the UK. Credit: Pexels

Microsoft wouldn’t say if it has any concerns about the way Uber is deploying its technology. A spokesperson told us that Microsoft is “committed to testing and improving Face API, paying special attention to fairness and its accuracy across demographic groups”. And that Microsoft “provides customers with detailed guidance” to “help them assess fairness in their system”. In a statement, Transport for London apologised to Imran Raja for any distress the removal of his license caused. It confirmed that Abdul Kadir Ali is pursuing an appeal against the decision to revoke his licence via the Magistrates Court. A spokesperson for TFL said: “The safety of the travelling public is our top priority and where we are notified of cases of driver identity fraud, we take immediate action to revoke a driver’s licence so that passenger safety is not compromised. “Where additional information is disclosed, for example relating to complaints raised against a driver, this also needs to be analysed before a decision to reinstate a licence can be made. We always try to ensure this is done in a timely way.”

In November 2019, TFL refused to renew Uber’s license to operate in London after 14,000 trips were made by 43 unauthorised drivers. The company later successfully appealed the decision in court.

Facial recognition technology is impressive but imperfect.

Those who have developed it are transparent about its capabilities, its limitations and the need to understand both if a system is to be fit for purpose.