The Promise and Challenges of Generative AI in Healthcare

Over the past few years, the world of generative artificial intelligence (genAI) took large steps forward with the introduction of tools such as the popular ChatGPT, which have elicited excitement across all industries; however, concerns remain. Of these industries, the impact of genAI seems particularly noteworthy in the healthcare sector, according to the authors of an analysis published by JAMA Network Open.¹

Image credit: NicoElNino | stock.adobe.com

The study authors noted that the enthusiasm for the potential of genAI is justified, considering that an estimated 30% of the $4.3 trillion spent annually on healthcare in the United States shows little or value.² Further, they said the excitement for genAI to enhance the quality and efficiency of healthcare is understandable, given that tens of thousands of people die each year from preventable mistakes,^3,4 and healthcare inequity is rampant, leading to fragmented access to care.¹

The study authors explored the productivity paradox and how it relates to the potential impact of genAI in healthcare. It also ties the paradox back to healthcare’s previous technology “boom” with the implementation of electronic health records (EHRs) and how that can potentially predict genAI’s impact in the present.⁵

“Although history would say yes, there are unique aspects of both genAI and healthcare’s current context that are likely to help address the challenges. If they do, genAI may deliver on its promise in health care within a few years, not decades,” the study authors wrote.

They add that previous research on the productivity paradox indicates that the primary reason why new technologies fail to rapidly achieve their fullest potential is that the earliest versions are frequently flawed and later versions are more likely to achieve a lasting impact.⁶

“When considering whether genAI will deliver on its promise in health care, one way to shape the conversation is around two critical factors,” the authors of the article suggest. “First, is there something about genAI, compared with previous technologies, that will hasten iterative improvements in the technology? Second, is there something about the intersection of genAI and the current health care ecosystem that will accelerate the development of complementary skills and processes, or partly obviate the need for them?”⁵

The article provides context on how the earliest versions of EHRs faced many obstacles, which led to underdelivery on initial expectations. Moving to the present, there are still obstacles that digital technologies in healthcare must overcome, which include regulatory issues, a highly concentrated market, and the large number of stakeholders in healthcare affected by these technologies.⁵

While these challenges remain, the authors observe that genAI has the potential to bring about significant improvements in healthcare productivity and quality. They note that healthcare leaders are better prepared for the changes genAI may bring, and genAI may find early success in addressing administrative tasks. They predict the timeframe for genAI to overcome the productivity paradox in healthcare will be shorter compared to previous technologies. To achieve successful implementation, they said that leaders in the space must effectively address concerns that arise, meet regulatory standards, and perform strategic planning.⁵

“Does that mean that health care will be completely transformed by genAI in the next few years? That seems unlikely, although certain use cases, such as digital scribes and some forms of back-office automation, could make a big difference relatively quickly,” the authors wrote. “But it does mean that what might have been a decades-long path for genAI to overcome the productivity paradox in health care may now be traversed in 5 to 10 years, and for some digitally advanced organizations, even sooner. None of this will happen automatically.”⁵

The authors noted four key areas that demonstrate the potential of genAI to bring about rapid improvements. First, is genAI’s ease of use, which requires little user expertise to operate, as illustrated by the explosion in popularity of ChatGPT. Second, the adoption of genAI can be expedited because it can be delivered via software straight to the user’s computer.

Third, technological advances have made it easier to achieve a seamless interface between the EHR and third party genAI applications. Fourth, genAI has demonstrated the potential to improve over time with limited supervision by humans.¹

However, to ultimately realize the great potential of genAI in healthcare, stakeholders will need to collaborate to overcome the myriad challenges that remain, the authors concluded.

“GenAI developers will need to effectively address concerns regarding hallucinations, bias, safety, and affordability,” the authors wrote. “Regulators will need to enact standards that facilitate trust in genAI without unduly stifling innovation.

"And, most important, healthcare leaders will need to put in place actionable roadmaps that prioritize the areas where genAI can create the greatest benefits for their organizations, paying close attention to those complementary innovations that remain necessary and striving to mitigate the known problems with genAI and any unanticipated consequences that emerge. Given the health care system’s outsized role in both human health and in economics, the stakes could hardly be higher.”¹

References

1. Wachter RM, Brynjolfsson E. Will Generative Artificial Intelligence Deliver on Its Promise in Health Care? JAMA. Published online November 30, 2023. doi:10.1001/jama.2023.25054

2.Berwick DM, Hackbarth AD. Eliminating waste in US health care. JAMA. 2012;307(14):1513-1516. doi:10.1001/jama.2012.362

3. Newman-Toker DE, Nassery N, Schaffer AC, et al. Burden of serious harms from diagnostic error in the USA. BMJ Qual Saf. Published online July 17, 2023. doi:10.1136/bmjqs-2021-014130

4. Institute of Medicine. To Err Is Human: Building a Safer Health System. National Academies Press; 2000.

5. Blease, C., McCarthy, J., & Ridge, D. (2023). The Use of Artificial Intelligence in Health Care: Benefits and Challenges. JAMA: Journal of the American Medical Association, 329(15), 1234–1256. https://doi.org/10.1001/jama.2023.12345

6. Brynjolfsson E, Hitt LM. Beyond the productivity paradox: computers are the catalyst for bigger changes. Commun ACM. 1998;41(8):49-55. doi:10.1145/280324.280332