The House Bipartisan Task Force on Artificial Intelligence closed out the year with a bang, launching 273-pages of AI policy fireworks.
The report includes recommendations to “advance America’s leadership in AI innovation” across multiple industries, and the healthcare section definitely packed a punch.
The task force started by highlighting AI’s potential across a long list of use cases, which could have been the tracklist for healthcare’s greatest hits of 2024:
- Drug Development – 300+ drug applications contained AI components this year.
- Ambient AI – Burnout is bad. Patient time is good.
- Diagnostics – AI can help cut down on $100B in annual costs tied to diagnostic errors.
- Population Health – Population-level data can feed models to improve various programs.
While many expect the Trump administration’s “AI Czar” David Sacks to take a less-is-more approach to AI regulation, the task force urged Congress to consider guardrails in key areas:
- Data Availability, Utility, and Quality
- Privacy and Cybersecurity
- Interoperability
- Transparency
- Liability
Several recommendations were offered to ensure these guardrails are effective, although the task force didn’t go as far as to prescribe specific regulations.
- The report suggested that Congress establish clear liability standards given that they can affect clinical-decision making (the risk of penalties may change whether a provider relies on their judgment or defers to an algorithm).
- Another common theme was to maintain robust support for healthcare research related to AI, which included more NIH funding since it’s “critical to maintaining U.S. leadership.”
The capstone recommendation – which was naturally well-received by the industry – was to support appropriate AI payment mechanisms without stifling innovation.
- CMS calculates reimbursements by accounting for physician time, acuity of care, and practice expenses, yet fails to adequately reimburse AI for impacting those metrics.
- The task force said there won’t be a “one size fits all” policy, so appropriate payment mechanisms should recognize AI’s impact across multiple technologies and settings (Ex. many AI use cases may fit into existing benefit categories or facility fees).
The Takeaway
AI arrived faster than policy makers could keep up, and it’ll be up to the incoming White House to get AI past its Wild West regulatory era without hobbling the pioneers driving the progress. One way or another, that’s a sign that AI is starting a new chapter, and we’re excited to see where the story goes in 2025.
