US govt wants black-belt AI experts, better biometric verification
The federal government is rolling out two incentive programs to push a variety of biometrics programs to evolve faster.
The first one is a series of industry challenges during 2023 that the Department of Homeland Security hoping will result in more secure, accurate and easier-to-use remote ID validation. The specific target is anyone trying to fraudulently apply for bank accounts and government services or create social media accounts.
The second effort, called a black-belt program, will try to identify biometric data champions.
The remote ID validation technology demonstration will be run by four Homeland Security segments: the department’s Investigations Forensic Lab, the Science and Technology Directorate, the Transportation Security Administration and the National Institute of Standards and Technology.
Officials want to get more understanding about digital identity security, privacy, equity and trust.
In a prepared statement, Arun Vemury, head of DHS’ Biometric and Identity Technology Center, acknowledges that market pressure has resulted in tools to identify applicants. But there are “questions about the performance and fairness of the technologies.”
Not only that, Vemury is concerned that people could exploit weaknesses in the tools and “commit fraud at scale.”
Two areas being scrutinized in this program involve assessing the liveness of selfie photos and verifying identities using images captured by mobile devices.
Separate from that, the government is looking at improved document authentication in this program as well.
The second program seeks Homeland Security’s AI experts in biometrics and fraud, to recognize them with a rating analogous to a black belt in martial arts.
According to reporting by trade publication Federal News Network, black belts will be sought throughout the department.
David Larrimore, chief technology officer of DHS, reportedly told the publication that recognized employees will “become a part of a larger community to help where people like you aren’t available.”
Larrimore noted the extent of the problem the federal government faces in covering its bases in biometrics.
He is quoted saying, “there is no way that, of the 350 or so acquisition programs going on right now, everyone has someone who could be considered an AI black belt.”
Larrimore is also looking for structural change.
Data has to be shared in a structured way. Black belts, it appears, would not be lone experts jumping in with their expertise and jumping back. The program will search for efficient ways of spreading what is learned.
Officials in the Defense Department and Postal Service, according to the article are ready to exploit the program to get the broadest possible AI literacy among employees and managers.
Ben Joseph, a departmental chief data officer within the Postal Service, told the publication that data literacy is a fundamental problem.
The best AI model is useless in the hands of people who cannot read a pie chart, said Joseph.