Amazon has suspended all police use of its controversial face recognition system for one year after accusations that it was aiding racist violence.
Dissident employees had charged the online shopping giant with hypocrisy for vocally supporting the Black Lives Matter movement while selling its Rekognition service to US police forces.
But the company announced on Wednesday that it would pause those deals while it waits for the US Congress to regulate face recognition technology, though it gave no specific reason.
Rekognition itself has been repeatedly found to be biased against black faces, with one study finding higher error rates than rival services from IBM and Microsoft. Amazon claimed the study was misleading.
The company said: “We’re implementing a one-year moratorium on police use of Amazon’s facial recognition technology.
“We’ve advocated that governments should put in place stronger regulations to govern the ethical use of facial recognition technology, and in recent days, Congress appears ready to take on this challenge.
“We hope this one-year moratorium might give Congress enough time to implement appropriate rules, and we stand ready to help if requested.”
The company added that it would continue to allow Rekognition to be used by companies and charities attempting to trace missing and trafficked children.
Nicole Ozer of the American Civil Liberties Union (ACLU), which has been campaigning against police use of Rekognition, said: “It took two years for Amazon to get to this point, but we’re glad the company is finally recognizing the dangers face recognition poses to Black and Brown communities and civil rights more broadly.
“This surveillance technology’s threat to our civil rights and civil liberties will not disappear in a year. Amazon must fully commit to a blanket moratorium on law enforcement use of face recognition until the dangers can be fully addressed, and it must press Congress and legislatures across the country to do the same.”
Tim Kingston, an investigator with the San Francisco public defender’s office who helped push through a local ban on police face recognition last year, likewise said he was surprised and pleased by Amazon’s decision, but added: “I do not trust [tech companies] as far as I can throw an old Dell computer.”
He attributed it partly to tech workers having “‘oh my God’ moments” as the cultural reckoning unleashed by the killing of George Floyd rippled out through Silicon Valley and Amazon’s native Seattle.
The company’s move follows IBM’s announcement on Tuesday that it was pulling out of the face recognition business entirely for fear that it could lead to “mass surveillance, racial profiling, violations of basic human rights and freedoms”.
Amazon’s shareholders previously shot down a ban on police use of Rekognition at this year’s annual meeting, with its board claiming that there had been no reports of the system being used “in the harmful manner posited in the proposal”.
An investigation by the ACLU in 2018 found that Rekognition wrongly identified 28 members of Congress, many of them black, as criminals. The company said the system had been calibrated wrongly.
A later study by the Massachusetts Institute of Technology, led by the crusading computer scientist Joy Buolamwini, found that Rekognition was wrong about the gender of dark-skinned women 31pc of the time, compared to an error rate of zero for white men.
Amazon disputed those results, citing internal research which it claimed had found “no significant difference in accuracy” across various ethnicities.
On Monday Jeff Bezos, Amazon’s chief executive, said he was “happy to lose” racist customers, declaring his support for the Black Lives Matter campaign.
His company’s statement on Wednesday did not mention its video doorbell division, Ring, which reportedly has partnerships with more than 500 US police forces to access its video footage but which has been accused of becoming a tool of suburban racism.