is urging governments world-wide to enact regulation of facial-recognition technology next year that requires independent assessment of accuracy and bias and prohibits ongoing surveillance of specific people without a court order.
Microsoft, which competes with both companies as well as
in the emerging market for facial-recognition products, but has been vocal in calling for government regulation of the technology, with a blog post in July arguing that change couldn’t happen if a few companies adopt new standards while rivals ignore them.
Brad Smith, Microsoft’s president and chief legal officer, dialed up the urgency on Thursday, arguing that delays to enacting new rules could “exacerbate societal issues.” Society is ill-served “by a commercial race to the bottom, with tech companies forced to choose between social responsibility and market success,” he wrote in a blog post.
Mr. Smith also was scheduled to speak about Microsoft’s position Thursday at the Brookings Institution in Washington, D.C., the same day a group of tech leaders from Microsoft and other companies visited the White House for a summit on issues including artificial intelligence.
Microsoft’s advocacy of regulation underlines the ambivalence over powerful new technologies enabled by advances in AI. Adoption of facial recognition is proceeding quickly—especially in China, where the government uses it extensively for surveillance—stirring concerns about potential misuse.
Mr. Smith declined to say if Microsoft would sell facial-recognition technology to China.
This year, dozens of civil-rights organizations called on Amazon to stop selling its facial-recognition technology to law-enforcement organizations. Asked about those concerns at a press conference last week, Andy Jassy, chief executive of Amazon Web Services, said the company hasn’t seen any abuses of the technology. He stopped short of calling for regulation, saying “countries themselves have to decide” on rules.
In his blog post, Mr. Smith listed benefits of facial-recognition including identifying missing children. Microsoft’s product, called Face, is used by customers such as Uber Technologies Inc., whose drivers take selfies to verify their identity when they launch the app to start picking up passengers.
But Mr. Smith highlighted three areas where governments should focus legislation: racial and gender bias, privacy and mass government surveillance. He cited George Orwell’s dystopian novel, “1984,” in which a government tracks citizens’ every movement, as a cautionary tale.
“Today technology makes that type of future possible,” Mr. Smith wrote.
He said that new laws should notify people—and get their consent—when facial-recognition technology is being used. And he said that governments need to continue to move quickly to address abuses that arise.
He acknowledged that Microsoft’s interests in regulating the technology are competitive as well.
“If a responsible company turns down business because it regards a particular use of facial recognition as likely to increase discrimination or abuse human rights, and then it sees its competitors go forward and gain those sales, you not only put people’s rights at risk, you risk tipping the market towards an approach that is less socially responsible,” Mr. Smith said in an interview.
Microsoft has discussed its ideas for legislation with both federal and state lawmakers, though Mr. Smith declined to name them. It also has talked about the need for regulation with its rivals and says he is “optimistic” they will join Microsoft.
Others calling for regulation of facial-recognition include rights advocacy groups such as the American Civil Liberties Union and the AI Now Institute. In a report published Thursday, AI Now echoed Microsoft’s calls for federal laws to provide oversight and transparency into the use of facial recognition.
Tech giants have adopted ethical principles around artificial intelligence, such as Google’s pledge earlier this year not to use AI in military weapons. Those principles were crafted to avoid missteps in developing new technology that could have harmful side effects. But they don’t go far in enough in holding large companies accountable, said AI Now co-founder Meredith Whittaker.
“It is great these companies are realizing they have a responsibility to ethical conduct,” said Ms. Whitaker, who also leads a research group at Google. But company-issued guidelines “are effectively promises by these corporations, on their own terms.”
AI Now, whose other co-founder is a Microsoft researcher, operates independently from those companies.