5 Reasons why AI is different in health care

©LALAKA-stock.adobe.com

For those of us who have been working closely with artificial intelligence (AI) for many years, it’s a little shocking to suddenly hear “AI” on the tips of everyone’s tongues. Although it may be difficult for the technology to match the current hype, there are many valuable use cases across industries, and that number is sure to grow quickly as the technology improves and organizations begin to experiment with different solutions.

Health care is no exception, but while AI undoubtedly will help clinicians become more efficient and improve patient outcomes, the sector differs from other fields in these important ways:

• ‘Good Enough’ is not good enough: In some fields, AI tools are going to perform slightly worse than humans, at least at first, and this is going to be okay. People are already using AI to draft response emails, for instance, and it’s not the end of the world if these tools fail to capture your tone and voice with perfect accuracy when confirming a video meeting. Basically, if an AI tool’s performance is only 90% as good as a human’s, but it makes a process significantly faster or simpler, that tradeoff will work for many people, workflows, and industries.

But this tradeoff won’t work in health care. Patients’ lives are on the line when clinicians change the way they deliver care, and providers simply won’t use AI tools that force them to compromise on quality—no matter how efficient those tools may be.

• The ‘Quintuple Aim’: Although some companies truly are committed to social equity, corporations have a fiduciary responsibility to maximize returns for their shareholders. This bottom-line focus stands in contrast to health care, where the idea of a “Triple Aim” – incorporating the patient experience, population health, and costs – has been widely accepted for many years. More recently, this has expanded to a “Quintuple Aim,” incorporating staff experience and health equity. The emphasis on equity, in particular, sets the goals of health care organizations apart from many businesses in other sectors. While corporations will often shutter stores in areas that have become unprofitable, health care has an ethical obligation to try to bring high-quality care to all populations, including in underserved locations. The AI tools adopted by the sector will reflect this emphasis.
• No one is losing their job: In some knowledge work fields, employees are growing nervous that AI will soon put them out of work. But the labor market in health care is so tight that it is nearly impossible to imagine health care organizations laying off clinicians after adopting AI tools. On the contrary, health care organizations are struggling with high turnover rates among staff who are burned out from mundane administrative tasks. AI tools will help clinicians complete this work, but there is little danger of the tools putting clinicians out ofwork.
• Vast quantities of data: Generative AI tools like ChatGPT have caused many people to rethink their previous assumptions about the limitations of the technology. These tools are now capable of producing writing and visual art that is, in some cases, difficult to distinguish from content created by humans, and many businesses are racing to incorporate them into their operations.

However, this rush is also making it clear that AI sometimes struggles with unstructured data. (ChatGPT has quickly become infamous for simply making things up.) By contrast, the vast quantities of population-level health care data generated by remote monitoring tools are a perfect fit for more established AI use cases like data analytics.

• Shared benefits of cost reductions: Even in nations with for-profit payers, such as the U.S., health care is essentially a societal cost. In other industries, organizations often benefit from technologies that increase demand and allow them to charge more. But no one really benefits from cost increases in health care. While other industries will look for ways to use AI tools to increase their profit margins, automation in the health care sector will be aimed almost entirely at improvements in quality and efficiency.

As AI tools take center stage, organizations across industries will experiment with emerging solutions, refine the most promising use cases, and eventually settle into a new normal where AI is a part of their everyday operations. But this journey will look different from sector to sector, and it is important that health care leaders carefully consider their unique challenges and opportunities as they seek to create a better future for both patients and providers.

Arnaud Rosier, M.D., Ph.D, is a cardiac electrophysiologist and founder and CEO of IMPLICITY®.