The Reason Big Tech Now Wants National Facial Recognition Laws

They fear a welter of state, local and municipal rules - and the possibility of an outright ban . By comparison, national laws look pretty good. JL


Ryan Tracy reports in the Wall Street Journal:

Amid rising calls for regulation, technology companies are pushing for laws that would restrict use of facial-recognition systems—and head off the more severe prohibitions some cities and states are weighing.Microsoft and and others stand to profit as government agencies and businesses expand use of the technology.“This technology is coming. What we should seek is a means by which to make sure that Big Brother is not coming”

Amid rising calls for regulation, technology companies are pushing for laws that would restrict use of facial-recognition systems—and head off the more severe prohibitions some cities and states are weighing.

Microsoft Corp., Amazon.com Inc. AMZN -2.24% and others stand to profit as government agencies and businesses expand use of the technology, which can require large investments in machine-learning and cloud-computing capacity.

That opportunity is threatened by campaigns to severely restrict its use.

San Francisco and six other cities have passed laws to block government use of facial recognition. Lawmakers in New York, Massachusetts, Hawaii and Michigan are considering some form of ban or strict limitation.

Pressed by advocacy groups, concert promoters LiveNation Entertainment Inc. and AEG Presents, which stages the Coachella Arts and Music Festival, say they don’t have plans to use facial recognition at their events.

More than 60 college campuses have also disavowed the technology, activists say—including the University of California, Los Angeles, which confirmed it nixed a proposal to link its security cameras to facial-recognition systems.

A coalition of 40 activist groups led by Fight for the Future is circulating “Ban Facial Recognition” petitions that call on lawmakers to block government agencies from any use of the technology. Erica Darragh of Students for Sensible Drug Policy, part of the coalition, says recruiting volunteers is a snap: “Facial recognition freaks people out.”

If elected president, Vermont Sen. Bernie Sanders says he would bar police from using it.

Against this backdrop, Microsoft is backing bills in Congress and in its home state of Washington permitting use of the technology with oversight.

“If we don’t move past the polarizing debates that have blocked progress, people will continue to be left without any protection under the law,” Microsoft President Brad Smith said in a statement.

The Washington state measures would allow facial recognition for specific uses such as investigating crime, controlling access to a building or identifying a ticket holder.

Amazon, the world’s largest provider of cloud-computing services, declined to comment, but has said it supports national standards. International Business Machines Corp. has called for “precision regulations” that don’t allow mass surveillance, and the CEO of Google owner Alphabet Inc. has said he is open to a temporary pause while regulations are developed.

Privacy advocates view industry-supported regulations as ploys to conduct business as usual.

“They are effectively geared to allow these companies to continue selling and profiting from these technologies, more or less unhindered,” said Meredith Whittaker of New York University’s AI Now Institute.

The market for technologies involving some form of facial recognition could be worth $14.5 billion in 2025, up from $2 billion last year, according to research firm Omdia.

Supporters see facial recognition as a means to keep intruders out of buildings, speed up entry lines at stadiums and airline gates, identify criminal suspects and locate missing children.

Opponents fear it will usher in a surveillance state. Participants in street rallies or public protests would lose their anonymity. Retailers could identify people entering their stores, possibly using it to monitor people with shoplifting convictions.

In the wrong hands, the technology could be used to target victims for financial scams, extortion or other schemes.

Studies also show that some facial-recognition systems are less accurate on nonwhite and female faces than on white males, although accuracy has been improving as the technology advances.

So far, law-enforcement agencies are among the biggest early adopters. One supplier, N.Y.-based Clearview AI Inc., says it has 2,000 active users at law-enforcement agencies, largely in the U.S. and Canada, including individuals who have been given a free trial.

New Jersey Attorney General Gurbir Grewal says Clearview’s facial-recognition app helped officers identify suspected sexual predators in an online sting operation led by Somerset County.

Amid rising concerns however, Mr. Grewal in January imposed a statewide moratorium on Clearview’s system until guidelines can be drafted. Even so, an outright ban would be “an overcorrection that could potentially undermine public safety,” he said in an interview.

Privacy advocates have raised concerns that Clearview hasn’t submitted its algorithm for federal accuracy testing, as other companies have, and that the free trials it gives to police agencies don’t go through the usual vetting for government contracts.

Critics also worry about Clearview’s methods. It scrapes internet photo databases, including social-media posts. Other systems run narrower searches, like scanning mug shots.

That broader dragnet raises the risk of abuse, critics say—enabling anyone with the Clearview app to take a photo of a stranger and potentially learn his or her identity and personal details.

Clearview CEO Hoan Ton-That says his firm compiles photos that are already public, and the technology is extremely accurate. Clearview limits use to law enforcement and security professionals, he said in an interview, though he acknowledged potential investors have been given access.

For police, Clearview is “only a lead, not a piece of evidence,” and they still need to prove their cases in court, he said.

In Washington state, a bill from Democratic state Sen. Joe Nguyen aims to bar police from using facial recognition for broad surveillance. They could deploy it to identify suspects or track people suspected of felonies but only after taking steps such as consulting the public and allowing third-party auditing.

“Technology itself is not what I’m scared of. I’m scared of how it’s used,” said Mr. Nguyen, who also is a senior program manager at Microsoft.

A second Washington state bill on consumer-data privacy requires companies to post notice when they use facial recognition in public. The operator of a facial-recognition system would have to get permission before storing someone’s likeness except when there is a “reasonable suspicion” the person was involved in a crime.

As the bills moved through the state Senate last month, Microsoft backed both. The Washington Association of Sheriffs & Police Chiefs opposed Mr. Nguyen’s bill, calling it overly burdensome. The American Civil Liberties Union criticized both bills for legitimizing surveillance, saying the state should let municipalities make their own choices.

The state House passed different versions of the bills and lawmakers are negotiating.

In Congress, members of both parties are discussing limits on federal use of the technology, including a pause on new uses, but haven’t reached consensus.

“This technology is coming. What we should seek is a means by which to make sure that Big Brother is not coming” along with it, said Rep. Clay Higgins (R., La.) at a Jan. 15 hearing of the House Oversight and Reform Committee.

One bill from Sens. Roy Blunt (R., Mo.) and Brian Schatz (D., Hawaii) would allow companies to use facial recognition with public notice, consent, and third-party testing. It has Microsoft’s backing, but hasn’t gained momentum.