Civil rights groups are calling on Amazon to permanently ban use of its facial recognition software, as an approaching deadline looms on the future of the technology.
In an open letter addressed to Amazon CEO Jeff Bezos and incoming CEO Andy Jassy, 44 civil rights groups pointed to ongoing instances of police violence against the Black community as evidence that Amazon should stop selling facial recognition technology to law enforcement. “As a company, Amazon has a choice to make: Will you continue to profit from selling surveillance technology to law enforcement? Or will you stand for Black lives and divest from giving law enforcement these harmful tools?” said the letter, which was published Monday.
After national protests that followed the death of George Floyd last year, Amazon followed Microsoft and IBM in stopping the sale of its facial recognition technology to law enforcement. However, unlike IBM, which abandoned its program, and Microsoft, which indefinitely suspended police use of its facial recognition until a federal law is introduced, Amazon opted to impose a one-year ban to “give Congress enough time to implement appropriate rules” to govern the use of the technology.
While some cities have imposed bans on facial recognition technology being used by police departments, the technology isn’t regulated by federal authorities. Amazon has yet to say whether it will continue its moratorium after it expires next month, or lift the ban and sell the technology to law enforcement.
“They did share that they are committed to standing with the Black community and standing for racial justice,” says Jennifer Lee, technology and liberty project manager at the ACLU in Washington State, where Amazon is headquartered. “If they’re going to do that they need to permanently divest from selling facial recognition technology and cease involvement with police and law enforcement.”
Amazon didn’t respond to requests for comment. However, the Seattle-based giant is pushing against shareholder calls for more transparency around the use of its facial recognition software, called Rekognition.
Ahead of the company’s annual general meeting on May 26, one shareholder proposal is calling for an independent third-party audit on the risks linked with government use of Rekognition, citing calls of more than 70 civil rights organizations to stop selling the technology, who said it contributed to “government surveillance infrastructure.” Another shareholder proposal is calling for an independent report on how Amazon conducts due diligence on its customers, including law enforcement agencies that use Rekognition.
In a proxy memo filed with the Securities and Exchange Commission, Amazon said that it has “conscientiously acted to review and address the concerns expressed in the proposal and transparently provided information regarding our actions to the public” and that it is actively engaged in policy debates around facial recognition regulation.
Amazon introduced Rekognition, a cloud-based technology that uses artificial intelligence and machine learning to identify people and objects in photos and video, in 2016. But the technology became a lightning rod for civil rights groups and anti-surveillance advocates after researchers at MIT found it identified gender of certain ethnicities less accurately than similar products made by Microsoft and IBM.
(Amazon said the MIT findings were “misleading and drawing on false conclusions” and asserted that its own tests had found no such inaccuracies.) After it was revealed the company pitched the software to the Immigration and Customs Enforcement agency, hundreds of Amazon employees sent an internal letter to CEO Jeff Bezos stating that they “refuse to contribute to tools that violate human rights.”
The heightened awareness around racial equality and concerns about police surveillance are making such shareholder proposals harder to ignore for institutional investors. Glass Lewis, a proxy advisory firm, issued a report last week recommending investors vote in favor of both shareholder proposals about Rekognition, given the previous controversies linked to the software, and the fact that no federal regulations appear set to pass before the moratorium passes.
“We have to draft these proposals in a way to get them on the ballot, so we go with a softer approach,” says Brianna Harrington, shareholder Advocacy Coordinator at Harrington Investments, which is bringing the proposal calling for an audit of risks linked to government use of Rekognition. “In a perfect world they’d stop selling the technology.”
I’m a staff reporter at Forbes covering tech companies. I previously reported for The Real Deal, where I covered WeWork, real estate tech startups and commercial real estate. As a freelancer, I’ve also written for The New York Times, Associated Press and other outlets. I’m a graduate of Columbia Journalism School, where I was a Toni Stabile Investigative Fellow. Before arriving in the U.S., I was a police reporter in Australia. Follow me on Twitter at @davidjeans2 and email me at firstname.lastname@example.org
In July 2018, the A.C.L.U. ran a study that it said matched the headshots of 28 members of Congress to mugshots of known criminals. A secondary test performed by the M.I.T. Media Lab in January 2019 and reported by The New York Times found that Recognition had a hard time identifying female faces and the faces of dark-skinned individuals. Representatives from Amazon, however, pushed back against those claims, saying both the A.C.L.U. and M.I.T. Media Lab studies didn’t use the Recognition technology properly.
The company also issued a lengthy response statement on how it uses Recognition. Lawmakers and other tech companies, though, are calling for greater oversight over the technology. The response to facial recognition Ahead of Amazon’s shareholders meeting, the San Francisco Board of Supervisors voted to ban the use of facial recognition technology by law enforcement groups, while Massachusetts currently has a bill seeking to put a moratorium on the tech in committee.
Microsoft (MSFT) President Brad Smith has said that his company rejected the sale of its own facial recognition technology to a police department out of fear that it would disproportionately impact women and minorities. Smith said that the technology had primarily been trained with white males, and, as a result, wouldn’t have been accurate. The company also denied the sale of its tech to a foreign country. Google (GOOG, GOOGL), meanwhile, has chosen not to sell its technology at all. For more on Yahoo Finance’s and Dan Howley’s coverage of this story please click: https://finance.yahoo.com/news/amazon…