Some members of Congress are asking the U.S. Health and Human Services Service to withdraw from years of efforts to create government-administered artificial intelligence control laboratories and create a model for an artificial intelligence control laboratory in cooperation with industry.
“We write to express our serious concerns about the potential role of assurance labs in regulatory oversight of artificial intelligence technologies and how this will lead to regulatory capture and stifling innovation,” Reps. Dan Crenshaw, R-Texas, Brett Guthrie, R- Ky., Jay Obernolte, R-Calif. and Dr. Mariannette Miller-Meeks, R-Iowa, said in a letter addressed to Micky Tripathi, acting chief AI officer at HHS.
WHY IT’S IMPORTANT
With deregulation a priority for the incoming Trump administration in 2025, Republicans are raising concerns about how artificial intelligence in health care will be controlled.
In a letter to Tripathi, who also serves as deputy secretary for technology policy and national health IT coordinator, representatives asked for clarification on the overarching goals of the agency’s reorganization, according to history on Monday.
As part of a larger technology restructuring effort by HHS, the modern ASTP — formerly the Office of the National Coordinator for Health Information Technology — announced in July that it would expand its responsibilities, including for artificial intelligence in health care, along with modern staff and more funding flowing to him.
The letter also questions ASTP/ONC’s statutory authority and role across the health care system through its creation of assurance labs to complement the U.S. Food and Drug Administration’s review of artificial intelligence tools, and suggests that there would be significant conflict interests.
“We are particularly concerned about the possible creation of paid assurance labs composed of competing companies,” the representatives said, adding that larger, incumbent technology companies could gain an unfair competitive advantage in the industry and negatively impact innovation.
Representatives asked eleven questions and requested answers by December 20.
An ASTP spokesman said via email that the agency could not comment on the letter at this time. CHAI did not respond to our request for comment, but this story will be updated if one becomes available.
A BIGGER TREND
One of the letter’s signatories, Rep. Miller-Meeks, had previously asked the then-director of FDA’s Center for Devices and Radiological Health about CHAI and its members.
During the House Energy and Commerce Health Subcommittee on the agency’s regulation of drugs, biologics and medical devices, Guthrie, as chairman of the subcommittee, said during opening remarks that several regulatory failures have created “uncertainty among innovators.”
Miller-Meeks specifically asked whether the FDA would task the coalition with certification. She noted that Google and Microsoft are founding membersand Mayo Clinic, which it says has deployed more than 200 AIs, employs some of the coalition’s leaders.
“It doesn’t pass the smell test,” she said, and shows “clear signs of trying to take control.”
CHAI, which unveiled healthcare AI transparency standards in line with ASTP’s healthcare IT certification requirements, said the long-awaited AI nutrition label will soon be released.
Earlier this year at HIMSS24, Dr. John Halamka, president of the Mayo Clinic Platform, addressed the significant potential benefits and real potential harms that could result from predictive and generative AI applied in clinical settings.
“Mayo has an assurance lab where we test commercial algorithms and self-developed algorithms,” he said in March.
“You just need to identify the error and then mitigate it. This can be mitigated by restoring the algorithm to different types of data, or simply understanding that the algorithm cannot be completely fair to all patients. You just need to be extremely careful where and how you use it.”
Since its founding in 2021, CHAI says it has been working to bring transparency to AI, create guidelines and guardrails to address algorithmic bias in health care, addressing government concerns and building on the White House AI Bill of Rights and Risk Management Framework AI and NIST support. Artificial intelligence assurance identified in President Joe Biden’s executive order on artificial intelligence directing HHS to establish a security program.
ON RECORDING
“The ongoing dialogue on artificial intelligence in health care must recognize the distinct powers and responsibilities of different agencies and offices to prevent overlaps that could lead to confusion among regulated entities,” the four Republican members of Congress wrote in their letter.