Facial Recognition's Latest Failure Is Keeping People From Accessing Their Unemployment Benefits
from the would-be-fixed-instantly-if-the-CEO-had-to-use-the-system-to-get-paid dept
Is there anything facial recognition tech can’t do?
The answer appears to be “yes.” Or maybe it’s “no.” The list of things it can’t do well flows from the only thing it’s asked to do and can’t: recognize faces.
It can’t allow you to board a plane at certain airports if it decides you’re not the person you actually are. It can’t keep you from being arrested if it decides you’re someone you’re not. And it can’t let you on the property if it’s deployed by any number of cameras watching any number of private establishments.
The latest thing it can’t do is keep unemployed people connected with their unemployment assistance. A private contractor acting as a fraud-fighting middleman is apparently fighting fraud by keeping legitimate recipients of assistance from receiving any assistance. Todd Feathers has more information at Motherboard.
Unemployment recipients have been complaining for months about the identity verification service ID.me, which uses a combination of biometric information and official documents to confirm that applicants are who they claim to be. The complaints reached another crescendo this week after Axios published a “deep dive” article about the threat of unemployment fraud based on statistics provided to the outlet by ID.me.
Some unemployment applicants have said that ID.me’s facial recognition models fail to properly identify them (generally speaking, facial recognition technology is notoriously less accurate for women and people of color). And after their applications were put on hold because their identity couldn’t be verified, many should-be beneficiaries have had to wait days or weeks to reach an ID.me “trusted referee” who could confirm what the technology couldn’t.
Months-worth of complaints can easily be found on Twitter. A steady stream of angry tweets in ID.me’s direction undercuts the company’s assurances that everything is working fine or, in the event that it isn’t, it’s all the end users’ fault.
In his statement to Motherboard, [IDme CEO Blake] Hall said that facial recognition failures are not a problem with the technology but with the people using it to verify their identity. “For example, if someone uploads a selfie that only shows half their face.”
“We are unaware of eligible individuals who have not been able to verify their identity with ID.me,” he said, and “the wait time for a live video chat session right now is less than five minutes and it has been consistently under 30 minutes all week.”
Hmm. Perhaps Hall should take a scroll through Twitter and try to square his statement with actual users’ experiences. Multiple complainants have noted the supposed 24/7 chat is anything but 24/7 and trying to reach someone by phone is even more futile than trying to utilize the seemingly all-but-abandoned online chat service.
While the backend goes neglected, CEO Blake Hall is pushing for more states to utilize ID.me to combat unemployment fraud. The service is currently used by 21 states but ID.me would obviously like that number to be higher. To push that number higher, Hall appears to be pushing other numbers higher. As Feathers points out in his report, Hall’s estimate of money lost to unemployment fraud leapt from $200 million in February to $400 million by the time he was approached by Axios earlier this month as it compiled its “deep dive” into the company. While this kind of growth is to be expected over time, the best source for fraud estimates isn’t a company that has a financial interest in portraying fraud as a bigger problem than it may actually be.
This is the latest frontier of facial recognition failure. And it will get worse before… well, there’s no way of telling if it will ever get better. More automation of services is inevitable. And efforts put in place to prevent abuse of those services will inevitably involve verification processes like these. But if the government is going to rely on third-party tech with this many known problems, it needs to make sure it contractors and its agencies have the people and processes in place to backstop AI decisions and head off more of the (also inevitable) failures.