On May 16, 2023, Sam Altman, CEO of OpenAI, testified before a Senate subcommittee regarding the potential for governmental regulation of artificial intelligence (AI). The hearing garnered significant media attention as Altman advocated for different types of regulation for AI. The proceedings were a stark contrast to other, more contentious congressional hearings involving tech company CEOs as lawmakers and witnesses largely agreed on the merits of policing AI in the future. Altman’s company, OpenAI, launched ChatGPT in November 2022 and grew its active user base to over 100 million within two months, making it one of the fastest-growing Internet applications to date. The AI chatbot can provide detailed responses to almost any user-generated prompt by running data from across the Internet through complex software known as a neural network.(1) The rapid growth and popularity of this technology and its many positive and negative impacts have prompted lawmakers in Washington to take notice. At the same time, healthcare technology companies are deploying it across the industry.
Earlier this year, Microsoft and Epic announced a collaboration that would integrate generative AI technology (ChatGPT) with EHR software to improve providers’ productivity. UC San Diego Health, UW Health (Madison, Wisconsin), and Stanford Health Care were among the first organizations to test this technology by allowing it to draft clinician responses to patient inquiries. A HIPAA-compliant version of ChatGPT was integrated with physicians’ inboxes in Epic’s MyChart patient portal to facilitate the responses.(2) These interactions are not meant to replace clinician interactions, but research does indicate that this technology can effectively engage patients under the right conditions. A recent study by JAMA Internal Medicine found that patients preferred ChatGPT responses to physician ones over nearly 200 questions posted in a social media forum. Evaluators preferred ChatGPT responses in approximately 79% of evaluations, chatbot responses were rated of higher quality than physician ones, and the physician responses were rated less empathetic than ChatGPT.(3) While the results of this study and the potential of this technology application are promising, leaders involved have emphasized the need for caution when relying on AI-generated responses.
Additionally, health systems are experimenting with AI to support various administrative tasks. Organizations like WVU Medicine (Morgantown, West Virginia) and LCMC Health (New Orleans, Louisiana) are using AI to enhance marketing activities. Both have reported using ChatGPT to assist in creating blog posts, analyzing market research, and drafting organizational documents such as job descriptions. Such efforts involve the software creating an initial draft and administrative team members providing revisions and edits before publication.(4) Other popular applications of AI in healthcare can be seen in revenue cycle areas such as denials management and collections.
The integration of AI technologies like ChatGPT in healthcare is gaining traction and demonstrating potential in various areas. However, amidst the rapid growth and promising results, caution is advised as leaders stress the importance of responsible implementation and the need for human oversight to ensure accuracy, empathy, and adherence to ethical standards. As AI continues to evolve and impact the healthcare landscape, it is crucial for policymakers, healthcare providers, and technology developers to collaborate and establish appropriate regulations that balance innovation with patient safety and privacy. As the advancements and corresponding regulatory debate(s) are progressing swiftly, look for additional developments in this space.
References
Chow A. How ChatGPT managed to grow faster than TikTok or Instagram. Time Magazine. https://time.com/6253615/chatgpt-fastest-growing/ . Accessed 5/30/2023.
Cerullo M. Question for your doctor? Artificial intelligence can help. CBS News. https://www.cbsnews.com/news/chatgpt-artificial-intelligence-ai-health-care-patient-messages/ . Accessed 5/28/2023.
Ayers JW, Poliak A, Dredze M, et al. Comparing physician and artificial intelligence chatbot responses to patient questions posted to a public social media forum. JAMA Intern Med. Published online April 28, 2023. https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2804309 . Accessed 5/28/2023. https://doi.org/10.1001/jamainternmed.2023.1838
Bruce G. The ‘potential is mind-blowing’: How hospitals use ChatGPT for marketing. Becker’s Health IT. https://www.beckershospitalreview.com/digital-marketing/the-potential-is-mind-blowing-how-hospitals-use-chatgpt-for-marketing.html . Accessed 5/30/2023.