Unions push for more Uber scrutiny after claims its photo ID tech was racist

A man checks his smartphone while standing amongst illuminated screens bearing the Uber logo
A man checks his smartphone while standing amongst illuminated screens bearing the Uber logo

Unions have called for a clampdown on how Uber uses automated systems after the company faced fresh accusations that its photo identification tech had been "racist" and led to some ethnic minority couriers being fired.

The Independent Workers' Union of Great Britain (IWGB) and GMB criticised Uber's automated systems, which include selfie technology for its couriers and drivers, so they can prove who they are to log onto the app.

Earlier this week, Wired reported that at least 14 British couriers for Uber Eats had had their accounts frozen or were fired after the systems were unable to recognise who they were. Uber Eats is understood to rely on both the automated systems as well as human verification, and argues accounts are never blocked purely because of AI software.

However, it led to a rebuke from IWGB that the system was "racist" with the union claiming it affected more people in the BAME community.

It is not the first time the criticism has been levied at identification systems, and US government researchlast year found that some facial recognition technologies were misidentifying ethnic minorities 10 to 100 times more often than white people.

In total, around 100 IWGB members had been fired as a result of decisions made by automated systems at courier companies, the union claimed.

In 2019, the Telegraph had reported that Uber faced a civil rights complaint over the termination of a black driver's account after its facial recognition software failed to recognise him.

Steve Garelick, a regional organiser for GMB, said: "Ultimately AI controls are not being controlled by anyone and I believe the ICO and government must put in strong controls."

IWGB hit out at Uber on Monday, saying the company was "governing people's lives through opaque algorithms". "If they are not even able to even do something as basic as recognise a person's face, how can we have faith that the thousands of terminations they make are correct? How can we trust that the pay is correct? Their responses to our demands to end unfair terminations have been as opaque as their algorithms."

It comes amid a campaign by IWGB to stop "unfair terminations" of workers across Uber and Deliveroo, which received the backing from a cross-party group of 60 MPs in November. Gig economy workers and their rights have been under the spotlight in recent weeks, with the Supreme Court ruling last month that Uber drivers should be given worker rights, such as holiday leave and sick pay.

That decision, whilst only applying to those workers who had filed the claim, is expected to kick off a whole host of other cases in the gig economy.

An Uber spokeswoman said: "Any decision to remove a partner from the Uber platform always involves a manual human review. In order to help prevent unauthorised users from using an account and to protect against potential fraud, drivers and couriers are regularly required to verify their identity through a Real-time ID verification check including before they can go online. Anyone who is removed can contact us to appeal the decision."