Credit: CC0 Public Domain

Washington state is getting ready to roll out the ID.me facial recognition system dropped by the IRS last month over privacy and equity concerns.

Human rights activists and some are asking state government officials, including in Washington, to ban the system. Amid the outcry, Washington's Employment Security Department says it doesn't use ID.me to verify the identities of people claiming jobless benefits—but plans to start doing so in June.

"Recently, we launched a competitive procurement process to secure a tool claimants would use to verify their identity before applying for benefits," said Nick Demerice, director of public affairs. ID.me, Demerice said, was the winning vendor.

Sold as a way to help stop scammers from accessing unemployment aid by adding an extra layer of security, ID.me offers government agencies and other customers a way to verify an individual's identity using information like a driver's license, Social Security number and a facial scan. The system prompts each user to take a "video selfie," holding the camera in front of their face to scan until they see a green check mark.

ID.me's is already used in 27 states' unemployment systems, as well as several federal agencies, including the Social Security Administration and the Department of Veterans Affairs. Four states credited the system with preventing $210 billion in fraud, ID.me says.

Washington launched a pilot project with ID.me in 2021, part of its efforts to combat a spike in fraud in spring 2020, Demerice said. Washington state officials found $1.6 million in fraudulent claims in March and April of that year, leading them to temporarily halt unemployment benefits.

The state did not implement the ID.me system following the pilot.

Recently, ID.me's facial recognition technology has come under fire for the amount of information it collects from applicants, which opponents of the system worry could be misused, resold or stolen.

Critics also say the system creates barriers that prevent people from accessing unemployment aid. The tech is not accessible to people who don't have smartphones, computers or . Facial recognition technology, they contend, also perpetuates racial biases.

A study from the National Institute of Standards and Technology found this type of tech led to more false positives for Asian and African American faces, compared with white faces, meaning it could have a disproportionate impact on people from marginalized communities through things like wrongful arrests.

Following backlash, the IRS in February backed off its plan to require anyone who wanted to access tax-related records online to record a video of their face with their computer or smartphone.

Now, activists are turning their attention toward other relying on ID.me's system.

Even as lawmakers pushed the IRS to drop ID.me for taxpayers, many were silent about the "tens of thousands of people" already impacted by the "discriminatory and harmful technology," said Caitlin Seeley George, a campaign director for digital rights nonprofit Fight for the Future.

"Many of our communities filing for unemployment across the country have felt demoralized and dehumanized by the ID.me process since the start of the pandemic," Seeley George continued. "Some were forced to wait for months to receive the lifesaving benefits they earned and helped to pay for. For these people, the dangers and harms of ID.me aren't theoretical—they're reality."

By the end of February, more than 20,000 people had signed a petition from Fight for the Future and 20 other groups calling for the end of government contracts with ID.me and to launch an investigation into the federal government's use of facial recognition technology.

Sens. Ron Wyden, D-Ore., Sherrod Brown, D-Ohio, and Elizabeth Warren, D-Mass., are pushing the U.S. Department of Labor to help state unemployment programs find a way to verify applicants' identities without relying on private contractors like ID.me.

Nearly 50 organizations signed a letter making a similar call, saying facial recognition technology broadly has been found to be biased and have a disproportionate impact on people of color and other marginalized communities.

The signing groups, which included the Consumer Federation of America, the Electronic Frontier Foundation and the NAACP Legal Defense and Education Fund, said there isn't enough data about ID.me's performance, including its face-verification failure rate and whether the system disproportionately fails to verify the identities of people of color.

It may also cut off access to jobless benefits for individuals with disabilities, including blindness, or applicants who may not have access to the necessary technology.

Blake Hall, founder and CEO of ID.me, said the system was created to increase access by offering more ways for individuals to verify their identity, particularly for people who do not have credit history, have little banking history, are without a home or who live overseas.

"Tying access to digital services to affluence is fundamentally unjust, and that's been the reality for access to government services in this country for decades," Hall said. "ID.me is the first company that's moved access rates up by untethering it from credit history. That's why I started the company."

Hall says ID.me follows the identity verification standards set by the federal government, and isn't in the business of collecting biometric data from individuals.

In response to concerns over its technology, the software company announced it will now allow users to verify their identity without using automated facial recognition and let them delete some identifiable information, starting March 1.

To safeguard against any bias with its facial recognition tech and speed up the identity verification process, ID.me also plans to introduce a new feature that allows a human to immediately review any individual's verification that failed.

In Washington, Demerice from ESD said the department has "noted the significant concerns raised about using facial recognition technology by ID.me in its work with the IRS."

"We take data security and data privacy very seriously," he said. "We will evaluate all information to ensure we implement this verification tool safely and responsibly. We understand the need to balance the competing priorities of claimant safety, ease of use and fraud prevention."

Demerice said in February the state was still in the process of determining how it would use ID.me, leading to confusion from groups like Fight for the Future and Washington's chapter of the ACLU that believed ESD was already using the system—in part because it was listed on ID.me's website.

ESD asked the company to remove the page that said Washington had "partnered with ID.me to verify claimants requesting benefits and services online," Demerice said.

But the confusion over whether Washingtonians applying for jobless benefits are required to consent to using facial recognition technology is part of the problem with this kind of tech, said Jennifer Lee, the technology and liberty project manager with ACLU of Washington.

"You can't outsource core government functions to a third-party vendor and not tell people how that system works, whether it has biases and even if it is using that system at all," Lee said. "We don't know the extent to which facial recognition technology is being used."