Many healthcare leaders would prefer that government regulators take a conservative approach to setting rules for artificial intelligence tools. One strong, common caveat exists, however: AI-generated content must be branded as such so anyone can see where it came from.
CEOs and high-level executives of healthcare-related organizations such as WebMD and Salesforce stressed the importance of innovation but accepted the growing role of boundary setting In a round-up of perspectives published by MobiHealth News.
In the past year, AI-enabled tools have proliferated throughout senior living and care settings, with new capabilities being added to remote monitoring or resident and patient health records to both collect data and provide new insights faster than human analysis.
AI tools also have been tested in clinical settings to see how accurately they can diagnose and predict several diseases that affect older adults, from heart disease to Alzheimer’s.
A high level of concern exists around biased, or inaccurate, results generated by AI, however, experts suggest. At the moment, AI tools are not able to self-correct when they train on biased data, and they don’t qualify responses that may be “hallucinations.”
Flagging content as AI-generated will allow for secondary review of accuracy, one interviewee suggested.
Over the past few months, congressional hearings have gotten the ball rolling on potential future laws about AI in healthcare, whereas federal agencies, and President Biden himself, have weighed in on the topic, balancing praise and caution.
The White House released an executive order on AI standards in October; the US Food and Drug Administration had announced an advisory committee on digital health technologies a few weeks prior.
In establishing patient safety and transparency rules, the government and health technology providers may be on the same page, suggested Brigid Bondoc, an FDA regulatory compliance lawyer, in a video interview.
“The government, and even global entities, are appropriately focused on issues like ethics and privacy,” Bondoc said. “Regulation is still tentative. Innovating in a way that helps patients is a shared goal of the FDA and industry.”
