As governments across the board aim to offer more and better services to citizens online, the process of verifying identity becomes more complicated, especially for transactions considered "high-stakes." At the NASCIO Midyear conference last April, Georgia Chief Technology Officer Steve Nichols outlined the scope of the challenge.
In 2015, Georgia applied for a grant from the National Strategy for Trusted Identities in Cyberspace (NSTIC) to pilot an approach that would help it cut down on tax fraud. As Nichols explains, the pilot program involves the Department of Revenue, the Department of Driver Services, MorphoTrust, and the central IT agency, which manages the infrastructure for the two state agencies involved.
How often does the system get it wrong, and fail to correctly link a taxpayer's selfie to their photo on record with the Department of Driver Services? According to a MorphoTrust representative, who spoke alongside Nichols on a panel at the event, the goal with facial recognition is to perform better than a person. Humans will make an error once in every 1,000 cases. MorphoTrust reported that they're on pace to meet their higher threshold of one error per 10,000 cases.
Government Technology editor Noelle Knell has more than 15 years of writing and editing experience, covering public projects, transportation, business and technology. A California native, she has worked in both state and local government, and is a graduate of the University of California, Davis, with majors in political science and American history. She can be reached via email and on Twitter.