White House AI policy a ‘seismic policy shift’ for biometrics developers and vendors

The World Privacy Forum (WPF) has logged its first impressions of the White House’s new AI policy, with a focus on biometric use cases. In language published as a blog post and a downloadable document, Executive Director Pam Dixon says the White House Office of Management and Budget (OMB) memorandum “provides an extensive and in some ways surprising articulation of emergent guardrails around modern AI.”

The OMB memorandum lays out minimum practices for “Safety-impacting and rights-impacting Artificial Intelligence.” According to the Forum’s analysis, guidance on biometric deployments is a noteworthy trend. “The biometric use case runs as a thread throughout the memorandum,” says the WPF. “For example, biometrics is expressly discussed in the memorandum’s section on responsible procurement. This section requires the government to assess when or if biometric information was collected without consent, was collected for other purposes originally, or was collected without validation of identity.”

In an explanatory fact sheet released in tandem with the OMB memorandum, the Transportation Security Administration (TSA) gets a special mention for its use of facial recognition in airports. The WPF underlines the White House’s statement that “when at the airport, travelers will have the ability to opt out from the use of TSA facial recognition without any delay or losing their place in line.” This, says the Forum, is “a clear call to every biometric developer, vendor, deployer, and end user that a seismic policy shift has taken place.”

“If biometrics will be used, nonconsensual training data will be a problem as will bias, untested or unvalidated systems. An initial and ongoing in-depth AI Impact Assessment is required.” Rules are rules, says the WPF, and this new set will cause significant disruptions in how biometrics providers conduct business across the U.S. federal government.

The WPF’s assessment comes off as mostly laudatory, in saying that “the memorandum’s statement of how biometrics should be deployed in Federal use cases is by far the most assertive, clear, and direct articulation of biometric governance by the U.S. government to date.” Indeed, “it may potentially be the most astute articulation of biometric governance in regards to AI-specific requirements and guardrails applicable to a national government anywhere to date.”

The authors of the EU AI Act might beg to differ. But, “regardless of extra-territorial comparisons,” the memorandum “sets a critically important and improved precedent in this aspect of AI governance, and will make inappropriate uses of biometrics by the U.S. Federal government or the procurement of unethically sourced or unverified biometric systems extremely difficult to support going forward.”

The White House has set a deadline of December 2024 for government agencies to have implemented the safeguards laid out in the memorandum.

Article Topics

AI  |  airport biometrics  |  biometrics  |  digital identity  |  facial recognition  |  U.S. AI policy  |  U.S. Government  |  World Privacy Forum