Artificial Intelligence & Machine Learning , Government , Industry Specific

White House Seeks Ideas on Boosting AI Privacy

The Biden administration is contemplating updating for the artificial intelligence age the privacy guidance that federal agencies must follow before activating new systems or adding a new collection of personal identifiable information to existing information technology systems.

See Also: Strengthen Cybersecurity with Zero Trust Principles

The Office of Management and Budget on Tuesday published a request for information that solicits input on the next generation of privacy impact statements - the decision tool federal agencies have used since 2002 to identify and mitigate privacy risks.

Among the questions the White House agency asks:

• What privacy risks specific to the training, evaluation or use of AI should agencies consider when conducting privacy impact assessments?
  

  The risks might include the inferences AI systems make about individuals, the White House said.

• What role should the assessments play in how agencies identify and report on their use of commercially available information - information obtained from the private sector that contains data about an individual or group of individuals?
• What privacy risks specific to commercially available information should agencies consider?

OMB also asked what improvements it should make to guidance for agencies about identifying and addressing privacy risks and whether agency privacy impact assessments should be easier to find and understand. Comments are due by April 1.

The impetus for the request for information comes from President Joe Biden's October 2023 AI executive order on AI, which instructs OMB to seek feedback on how to improve privacy impact assessments, especially in light of privacy risks "further exacerbated by AI."

The White House on Monday touted progress in meeting milestones set out by the order (see: Biden's AI Executive Order, 90 Days On).

Concerns about technology-driven privacy risks have existed for decades, but AI has given them new urgency. In October, members of the federal legislative branch involved in a yearslong debate over a national privacy law tied AI privacy concerns to lack of a national privacy standard (see: US House Panel: AI Regulation Begins With Privacy).

Congressional members have vowed to pass federal privacy legislation this year, but election years shorten the number of legislative working days available and often push issues such as privacy to the margins.