US hospitals are deploying AI fast — but oversight is dangerously lagging

Full text

Walk into almost any U.S. hospital today and you will find AI doing some of the important work of medicine: drafting clinical notes, flagging sepsis, screening imaging, conducting prior authorizations, and answering patient messages. Adoption is moving fast, and the benefits are real.

But there’s a problem. Most health systems are monitoring the safety and performance of these tools the same way they govern a new MRI scanner in 2010: a subcommittee, a checklist, a quarterly meeting, an approval or a rejection. The process could take six months or longer.

That approach was already strained for older digital tools. For AI, it is dangerously inadequate, and the responsibility for fixing it rests not with IT, but with senior leadership.

Continue to STAT+ to read the full story…