The letter follows several recent Capitol Hill hearings and GAO reports raising concerns about the growing use of such technologies by federal agencies. Image: MONOPOLY919/Shutterstock.com
By: FEDweek StaffA group of 16 Democratic House and Senate members has pressed the White House to incorporate strong controls over use of artificial intelligence by federal agencies in a planned upcoming executive order.
A letter to President Biden recommends that the order build on and expand principles in an earlier “Blueprint for an AI Bill of Rights” from the administration, saying that applications such as facial recognition technology already has led to wrongful arrests and wrongful denials of benefits.
They recommended that the order include requirements for agencies “including equity and bias assessments, ongoing monitoring, data minimization practices, public consultation around the development and deployment of AI systems, and public disclosures about the implementation of these principles. Moreover, although the AI Bill of Rights did not apply in the law enforcement or national security contexts, the AI executive order should clarify that, wherever possible, these principles should apply across the federal government,” they wrote.
“In particular, these principles should apply when a federal agency develops, deploys, purchases, funds, or regulates the use of automated systems that could meaningfully impact the public’s rights. For example, if an agency creates an algorithm for internal use, it would need to undergo an algorithmic impact assessment by an independent assessor. Similarly, an agency seeking to purchase a service involving an algorithm would need to require vendors to complete a pre-purchase independent algorithmic impact assessment,” they wrote.
The letter follows several recent Capitol Hill hearings and GAO reports raising concerns about the growing use of such technologies by federal agencies.
A bill (S-1865) ready for a full Senate vote would require agencies to notify individuals when they are interacting with or subject to decisions made using certain AI or other automated systems. It also would require agencies to establish an appeals process that will ensure there is a human review of AI-generated decisions that may negatively affect individuals.
Nearly 10,000 Federal Offices Don’t Meet Usage Standards
OPM Plan on Employee Ratings Asking for Abuse, Says Senior House Democrat
OK, FERS and TSP, but What About Social Security Retirement Income?
Conversions to Schedule P/C Pending; Acknowledgement Form Draws Attention
Senate Passes DHS Funding Deal, but Stalls in House; Trump Signs Order to Pay TSA Personnel
See also,
Calculating Service Credit for Sick Leave At Retirement
FERS Supplement vs The 10% Pension Bonus
How Your FERS, Social Security and TSP Payments Get Taxed
