Rank One highlights biometrics supply chain risk to U.S. national security
Supply chain challenges in tech and other industries have been making headlines for well over a year, but the security risk to supply chains for artificial intelligence and machine learning systems, such as for facial recognition, are drawing scrutiny from a leading provider of biometric algorithms to the U.S. federal government.
Rank One Computing CEO B. Scott Swann addressed the issue at Connect:ID in Washington, DC in early October, suggesting that systems playing critical roles in national security should be free from the “limitations and vulnerabilities that foreign technology components can introduce.”
The company says that following on an executive order by President Biden earlier this year on domestic procurement and the wave of high-profile data breaches, it is working with the industry to mitigate supply chain risk to the national identity screening ecosystem.
ROC Chief Scientist Dr. Brendan Klare tells Biometric Update in an email that the main concern for the supply chain is with hidden software behaviors that can compromise AI and ML systems.
“Using foreign biometrics technology could mean that certain persons are predisposed not to be identified by the system. This could be, for example, members of a nation’s clandestine service, who can easily enter other countries without being matched to their true identity. Other concerns are backdoors which allow biometric information to be siphoned out,” Klare says, referring to the reported appearance of code written by a Russian government-linked entity in FBI software years ago.
The concerns Rank One has highlighted apply mostly to software, but Klare notes that the U.S. is also moving towards independence in semiconductors.
When asked what the short-term future holds for the U.S. national identity ecosystem’s supply chain, Klare says, “More and more focus on U.S. owned and developed technologies, not just for U.S. government, but also private enterprise. The risks are too high given the known cases of abuse.”
AI | biometric identification | biometrics | machine learning | national security | Rank One Computing | U.S. Government