PhotoLate last week, Amazon CEO Jeff Bezos received a letter from shareholders who are concerned about the privacy threat of government surveillance from facial recognition technology.

Rekognition -- Amazon’s technology -- was introduced in 2016 and can detect objects and faces in images and videos. It also allows people to upload face databases to automatically identify individuals, technology that is utilized by law enforcement officials in Orlando, Florida and Washington County, Oregon.

Now, shareholders including the Social Equity Group and Northwest Coalition for Responsible Investment are joining groups such as the American Civil Liberties Union (ACLU) to stop Amazon from selling the software. The groups cite the risks of mass surveillance.

“We are concerned the technology would be used to unfairly and disproportionately target people of color,  immigrants, and civil society organizations,” the shareholders wrote. “We are concerned sales may be expanded to foreign governments, including authoritarian regimes.”

The groups go on to write about potential civil and human rights violations, and how this could potentially hurt the company’s stock in the future. Additionally, they reference Facebook’s recent controversies as a warning of what could happen to Amazon.

Amazon’s response

Despite shareholder concerns, Amazon believes cutting off law enforcement to this technology would limit the power and scope of this and future technologies. However, the company policy also prohibits the use of its service for activities that are illegal, violates the rights of others, or may be harmful.

“We believe it is the wrong approach to impose a ban on promising new technologies because they might be used by bad actors for nefarious purposes for the future,” said Matt Wood, a general manager of artificial intelligence at Amazon Web Services.

“The world would be a very different place if we had restricted people from buying computers because it was possible to use that computer to do harm.”

A look into Rekognition

In April, Amazon announced that Rekognition is now up to 25 percent more accurate in picking out faces than its previous versions; however, it’s still unclear how accurate the technology is.

“Currently, there are no best practices for accuracy for facial recognition, which should be quite troubling to everyone,” said University of District Columbia law professor Andrew Guthrie Ferguson. “Accuracy is a hard issue to benchmark. As a society, are we okay with a 50 percent false positive rate, or a 20 percent false positive rate, when it comes to stops, arrests, or police investigation? The answer to the accuracy question will determine who gets handcuffed and who does not.”

“It’s been like pulling teeth to get information from governments or Amazon about what they’re doing to ensure this product does not harm people living in communities across America,” said ACLU attorney Matt Cagle. “We know very little about how accurate or how useful Amazon’s system really is.”

Overall, the shareholders and experts in the technology field are concerned about the implications this could have moving forward.

“The implications of Amazon Rekognition and all new facial recognition technologies is nothing less than a rebalancing of power between citizens and the police,” Ferguson said. “The ability to identify, track, and monitor everyone throughout the city is something that we read about in science fiction.”


Share your Comments