ID Tech Digest – October 15, 2024

Welcome to ID Tech’s digest of identity industry news. Here’s what you need to know about the world of digital identity and biometrics today:

Neglected ‘Alpha’ Channel Opens Door to Attacks on Computer Vision Systems: UTSA Researchers

Researchers from the University of Texas at San Antonio (UTSA) have identified a major vulnerability in artificial intelligence image recognition systems, including facial recognition. Led by Associate Professor Guenevere Chen and Qi Xia, the study shows that many AI models fail to process the alpha channel, which manages image transparency. This oversight allows for the “AlphaDog” attack, which manipulates images in ways imperceptible to humans but can trick AI systems. The vulnerability poses risks in various sectors, including facial recognition for security, autonomous driving, and medical imaging, where manipulated data could lead to misidentifications or faulty outcomes. The team is working with major tech companies to address this gap by incorporating full alpha channel processing in AI systems.

UK Fintech Firm Implements Random Selfie Checks to Fight Fraud

UK fintech firm ANNA Money has introduced a new biometric re-authentication feature that requires customers to verify their identity with a selfie when accessing accounts from unfamiliar devices. To enhance security further, the company has also implemented random selfie checks to prevent unauthorized access and combat financial fraud, such as account takeovers and money laundering. If the biometric data does not match the account holder’s original profile, the account is suspended. The security measure aligns with the UK’s new rules requiring financial institutions to reimburse victims of Authorised Push Payment (APP) fraud.

Luxand Marks NIST Testing Success, Launches Upgraded SDK

Luxand, Inc. has announced the successful performance of its FaceSDK in the National Institute of Standards and Technology (NIST) Face Recognition Vendor Test (FRVT), where it placed within the top 21.8 percent in the Visa-Border Combination test. The test evaluates the effectiveness of face recognition systems in real-world conditions, comparing high-quality visa photos with less ideal images taken at border crossings. Luxand has also launched version 8.2 of FaceSDK, introducing an iBeta Certified Liveness Add-on to counter spoofing attempts and offering enhanced support for platforms like Android, iOS, Windows, and Linux.

ID.me Share Sales May Point to IPO: Report

ID.me is reportedly offering its employees and early investors the chance to sell shares in a deal that values the company at $1.8 billion, a slight increase from its prior valuation of $1.73 billion. The tender offer, which provides $67 million for purchasing shares from existing stakeholders, allows the company to give liquidity to early investors without diluting ownership or raising new capital. The offer involves prominent investors like Ribbit Capital, Viking Global, and CapitalG, and has fueled speculation of a potential IPO in 2025. ID.me gained attention for its role in government services during the COVID-19 pandemic but has faced scrutiny over its fraud claims and use of facial recognition technology, especially in its partnership with the IRS.

iDenfy Extends Reach in Crypto Market with GlolinkOTC Partnership

iDenfy has partnered with GlolinkOTC, a Czech fiat-to-crypto exchange, to integrate its identity verification and compliance tools into GlolinkOTC’s platform. The collaboration will enhance the exchange’s Know Your Customer (KYC) and Anti-Money Laundering (AML) protocols, providing real-time AML screening, politically exposed persons (PEP) checks, and adverse media monitoring to meet regulatory standards and mitigate fraud risks. iDenfy’s platform includes AI-driven facial recognition and liveness detection, ensuring secure and efficient onboarding. The partnership aims to bolster GlolinkOTC’s security framework while maintaining user convenience.

TSA Brings CAT-2 Kiosks to Montana

The Transportation Security Administration (TSA) has deployed its latest Credential Authentication Technology (CAT-2) with facial recognition capabilities at five major airports in Montana: Bozeman, Billings, Great Falls, Helena, and Missoula. These CAT-2 units enhance the identity verification process by matching a live photo of the traveler with the image on their ID, offering a more secure and efficient method of confirming traveler identities. Unlike the earlier CAT units, which required TSA officers to visually confirm the match, CAT-2 automates this process with facial matching technology. Participation in the facial matching process is optional, and the TSA emphasizes that the photos are used only for immediate verification and are not stored.

No Bias Found in New Zealand’s Biometric Identity Check System, Clearing Path for Wider use

New Zealand’s government has confirmed that its Identity Check facial recognition system, in use by the Ministry of Social Development (MSD), has no significant racial bias following extensive testing. The system allows users to verify their identity by matching a smartphone photo to their passport or driver’s license photo. Independent testing by Australia’s BixeLab, an accredited lab, showed that the system performs consistently across New Zealand’s diverse ethnic groups, including Māori, Pasifika, and Pākehā populations. This milestone clears the way for broader use of the system, which has already been adopted by government and organizations like Hospitality New Zealand. Plans are in place to expand its use to sectors like banking and education, enhancing efficiency and accessibility for identity verification.

Dubai Cuts Cruise Ship Immigration Screening by 20 Mins with Facial Recognition Tech

Dubai has introduced a facial recognition system for cruise ship arrivals, reducing immigration processing time to around 20 minutes. The new system, called Portable Marina, allows immigration officers to pre-process passenger information digitally before the ship docks. Upon disembarkation, facial recognition technology verifies passengers’ identities quickly, enabling the system to handle up to 5,000 passengers in just 20 minutes. This initiative is part of Dubai’s ongoing efforts to streamline immigration, building on earlier implementations such as biometric e-gates introduced in 2011 at Dubai International Airport. The vendor behind the Portable Marina system has not yet been confirmed, though Vision-Box, Emaratech, and Idemia have previously provided biometric technologies for the UAE’s border control projects.

Mastercard Enters Age Verification Arena

Mastercard has launched a new “Identity Attribute Verification” service, aimed primarily at age verification for online transactions. The service allows merchants to confirm customer details such as age, date of birth, or address through their card issuer, facilitating purchases that require age validation, such as alcohol. This move comes as global regulations tighten around verifying minors’ access to certain products and services, with states in the U.S. and the UK implementing age check requirements for social media and adult content. Mastercard plans a phased rollout, beginning in select European countries and expanding in 2025.

House Science Committee Wants Answers from NIST on Facial Recognition Concerns

Three House Republicans from the House Committee on Science, Space, and Technology have sent a letter to the National Institute of Standards and Technology (NIST) expressing concerns over the agency’s role in facial recognition technology guidelines for federal online services. Committee Chair Frank Lucas, along with Representatives Mike Collins and Jay Obernolte, are seeking clarification on how NIST ensures privacy, accuracy, and safeguards around the use of biometric data. Their concerns focus on the reliability of facial recognition across diverse demographics and the protection of personally identifiable information (PII). The inquiry follows broader congressional concerns regarding AI and biometric technologies, with lawmakers urging stronger federal oversight to prevent algorithmic discrimination. NIST has acknowledged receipt of the letter and is preparing a response.

—

October 15, 2024 — by Tony Bitzionis and Alex Perala