FB pixel

Socure pitches biometric ID verification to close an underaged drinking loophole

Socure pitches biometric ID verification to close an underaged drinking loophole

Socure is using biometric ID verification to stop those who are underage from accessing alcohol through delivery apps that require little to no age verification.

Back in May, GoPuff was accused of 19 counts of selling alcohol to minors in the Boston College area, resulting in the company losing its license to deliver alcohol. In some cases, students presented fake IDs, while in others, they weren’t asked for identification at all.

Some fake IDs are virtually indistinguishable from their legitimate counterparts and can successfully pass through an ID check. More obvious fakes put the delivery driver in the awkward position of determining how to respond when they come across a fake.

“Are they going to complete the delivery? Should they be confiscating the fake ID?” asks Eric Levine, head of Document Verification at Socure, in an interview with Biometric Update. These questions could “result in drivers not wanting to check IDs to avoid the challenging situation that they’re put into.”

Socure’s Predictive DocV uses a multilayered approach to document verification that can be incorporated into the apps themselves, taking the onus off of drivers to suss out fake IDs in person, which UK research into retail age checks makes clear could jeopardize their safety.

“The delivery driver’s job becomes much easier. The hard part of detecting that fake ID, and frankly, stopping fake IDs before they’re used has been achieved and at this point.”

Socure’s document verification uses predictive machine learning fraud models that are trained from over 400 third-party data sources and uses Socure’s graph of 1 billion known good and bad identities, as well as risk signals like geolocation, behavioral data, and device ownership, among others.

At the end of June 2023, Socure acquired Berbix, the developer of a forensics and data extraction engine that reads OCRs, MRZs and barcodes through image processing to identify compromised and fake IDs.

Socure’s most recent version of the document verification software – DocV 3.0 – utilizes both Berbix and Socure’s software.

The company says DocV 3.0 can accurately capture 83 percent of all fraudulent documents compared to the industry standard of 56 percent.

First, the user takes a picture of the front and back of their ID along with a selfie. The software automatically compares face biometrics from the images to confirm they are the same person. It also uses image-of-image detection to identify biometric presentation attacks in the form of recaptured images, which is one of the common types of fraud attempts Socure comes across, says Levine.

It also looks for inconsistencies in the data in barcodes on the back of the card along with predictive analysis to identify signs of a fake ID.

“Maybe it’s a fake address that doesn’t actually exist in the real world,” says Levine. “By leveraging Socure’s identity graph, we can look for consistencies or inconsistencies to give us the highest likelihood of detecting that something is not owned by that particular individual or it doesn’t exist in the real world.”

From a customer experience perspective, an alcohol delivery company might not want to use document verification that makes it challenging for customers to input usable images or takes valuable minutes to process.

DocV3.0 coaches users to submit usable images of IDs and selfies and can determine whether to reject or accept a credential in under 4 seconds. “So if a user is going through that process,” Levine says, “they can get back to what they were doing, ordering alcohol, without any undue delay, to make sure that it’s as good of a user experience as possible.”

Related Posts

Article Topics

 |   |   |   |   | 

Latest Biometrics News


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Most Read This Week

Featured Company

Biometrics Research

Biometrics White Papers

Biometrics Events

Explaining Biometrics