What’s in Schumer’s AI framework for health care

Senate Majority Leader Chuck Schumer’s bipartisan artificial intelligence working group is calling for committees to consider a policy framework that could rein in the technology’s use in health care.

The policy proposals, from the working group of Schumer and Sens. Mike Rounds (R-S.D.), Martin Heinrich (D-N.M.) and Todd Young (R-Ind.), are the most detailed Congress has offered on AI in health care to date. Lawmakers have put forward individual bills on issues like biosecurity and bolstering research, but there hasn’t been a broader vision from leadership. The group put together the proposal after holding a number of forums with top AI industry officials, academics, industry groups and more.

Legislators hope to harness AI to shake up the health care system while mitigating its potential risks, including bias and privacy issues. Health industry groups are calling for regulation of the technology, though pushing for a lighter touch.

Schumer and the group say they expect lawmakers to work through the committee process and not wait for a larger legislative package to materialize, but rather move individual bills once a consensus is reached. That could be a tall order in a historically unproductive Congress, especially ahead of the elections in November.

The group also backs a broader federal data privacy framework.

Here are some of the most significant proposals:

High-risk AI: Schumer and his colleagues called for committees to “support efforts” on a “risk-based approach” to AI, meaning higher-risk applications would get more scrutiny.

The group called on committees to scrutinize guidance from agencies on “high impact” AI, which could include health care use, and weigh whether requirements are needed for users to understand why AI tools came to a certain conclusion. It also called for lawmakers to explore whether some uses of the technology should be “extremely limited or banned.”

Guardrails: Schumer and the bipartisan group called for lawmakers to weigh legislation that could support “further development of AI” in the sector but also have “appropriate guardrails and safety measures” for patients, including consumer protection, fraud prevention and ensuring “accurate and representative data.”

The lawmakers called for consideration of transparency requirements for providers and the public for medical products and clinical support technology, as well as the underlying data.

Payment: The four senators said lawmakers should explore CMS’ payment methods for AI, as well as “guardrails to ensure accountability, appropriate use, and broad application of AI across all populations.”

Research: The working group said it hopes the executive branch and appropriators can work together to commit at least $32 billion in non-defense AI spending annually. That would include a “cross-government” research and development endeavor including the National Institutes of Health and a focus on AI that would “fundamentally transform” medicine.

The report called for NIH to focus on data governance and making health data available for research.

Biosecurity: The technology has the potential to heighten bioweapon-related risk, the lawmakers said.

“The AI Working Group encourages the relevant committees to consider the recommendations of the National Security Commission on Emerging Biotechnology and the NSCAI in this domain, including as they relate to preventing adversaries from procuring necessary capabilities in furtherance of an AI-enhanced bioweapon program,” they wrote.

Drug crackdown: The lawmakers called on committees to craft legislation using “modern data analytics” to combat drug trafficking, including fentanyl and other synthetic opioids.

Liability: The group called on committees to weigh if there should be more standards surrounding holding AI users and developers accountable if the technology causes harm.