Amazon said it's joining a push for legislation on facial recognition technology.
The online retail has been offering its facial recognition software, called Rekognition, for the past two years. In that time it has become asurrounding the tech because it provides Rekognition to some law enforcement agencies in the US.
Michael Punke, Amazon Web Services' vice president of global public policy, suggested in a blog post Thursday guidelines for potential legislation or rules around facial recognition software. His comments follow repeated calls by fellow tech giant Microsoft, which also develops facial recognition tech, to create legislation for the software.
"New technology should not be banned or condemned because of its potential misuse," Punke said in the post. "Instead, there should be open, honest and earnest dialogue among all parties involved to ensure that the technology is applied appropriately and is continuously enhanced."
Amazon's statements offer an apparent softening of its position. Until now, the company has often defended its work on facial recognition when challenged, mentioning its value in finding missing children and more quickly solving crimes. The company has faced a constant barrage of criticism for the technology, with civil rights groups led by the American Civil Liberties Union raising concerns about the potential for misuse of Rekognition, excessive surveillance and bias against certain races or groups.
In New York City, where Amazon is planning to build a major new headquarters, city council members have often called out Amazon's marketing of Rekognition to US Immigration and Customs Enforcement, a business connection Amazon has declined to confirm exists.
In less controversial settings, facial recognition tech is also used to unlock Apple iPhones using the Face ID feature and to check travelers' identities at government kiosks at US airports.
In Thursday's blog post, Punke said the company came up with a legislative outline after discussing facial recognition with customers, researchers and policymakers over several months. The technology shouldn't be used to make fully automated and final decisions when used in law enforcement, he said, and should instead require a human to review the results. In cases of policing, Punke said, authorities should use a 99 percent confidence score, which is a measure of how sure a facial recognition system is that it found a match.
Also, he said, there should be a notice posted if video surveillance is combined with facial recognition in any public or commercial setting. In several states, that combination is already illegal.
"Our communities are safer and better equipped to help in emergencies when we have the latest technology, including facial recognition technology, in our toolkit," Punke said.
In response to Amazon's post, Neema Singh Guliani, ACLU's senior legislative counsel, disputed several of Amazon's claims. For example, she said a 99 percent threshold "does nothing" to prevent law enforcement from using facial recognition from keeping tabs on public protestors, immigrants and communities of color.
"Amazon's framework rings woefully hollow, underscores the company's refusal to properly address the dangers of its technology in government hands," Guliani said, "and reinforces the urgent need for Amazon to get out of the surveillance business altogether."
First published at 10:20 a.m. PT.
Updated at 11:45 a.m. PT: Adds comments from ACLU.
Taking It to Extremes: Mix insane situations -- erupting volcanoes, nuclear meltdowns, 30-foot waves -- with everyday tech. Here's what happens.
Fight the Power: Take a look at who's transforming the way we think about energy.