By Michael Balas.

In the swiftly advancing realm of artificial intelligence (AI), a tantalizing question emerges: can AI systems help us navigate the murky waters of medical ethics? Our recent study, published in the Journal of Medical Ethics (JME), tackled this very question, and our findings were enlightening.

At the heart of our research was OpenAI’s GPT-4, a sophisticated language-based AI, tested for its capability in the intricate domain of medical ethics. Why did we choose this area of inquiry? Medical ethics grapples with life-altering decisions and involves a complex interplay of stakeholders. If AI could assist here, its implications for healthcare would be monumental.

The results were a mix of promise and caution. GPT-4 demonstrated an impressive ability to identify major ethical concerns and provide procedural guidance. This was no minor feat; it showed that there is potential for AI to act as an auxiliary tool for medical ethicists, assisting in structuring discussions or pinpointing vital ethical factors.

However, while GPT-4 was able to recognize and apply foundational medical ethical principles such as beneficence and autonomy, it was not without its flaws. The AI, as anticipated, exhibited a systematic and mechanistic approach to ethical dilemmas, frequently neglecting the subtle human elements inherent to these situations. For instance, it sometimes equated an individual’s emotional state directly to their decision-making capacity. Such inconsistencies highlight the potential for misguided ethical advice and underline the crucial role of human oversight.

The impetus for our study stemmed from two primary concerns. First, there was a genuine intrigue surrounding the nexus between AI’s computational prowess and the complexities of human decision-making. As AI technologies become increasingly integrated into various aspects of our lives, it is essential to grasp how they might interact with, influence, or even replicate the multifaceted cognitive processes humans employ when making choices.

Further intensifying our research interest is the rapid proliferation of AI across sectors. With such technologies becoming ever-present, from our smartphones to our hospitals, gaining a comprehensive understanding of their capabilities becomes not just beneficial but vital. This is especially true in areas that demand a high degree of sensitivity and nuance, like medical ethics. As AI systems take on roles in these delicate realms, it becomes critically important for us to discern not only what they can achieve but also where they might falter or misinterpret the deeply ingrained values and principles that guide human ethical considerations.

Our study is not exhaustive. We analysed a limited number of ethical dilemmas and relied on a select group of ethicists for evaluations. Yet, it serves as a starting point for a conversation that is crucial in today’s digital age. It’s an invitation for more dialogue between AI developers, ethicists, healthcare professionals, and the public.

To those curious about the intricacies of our research, I invite you to delve into our full paper in JME. For now, I leave you with this thought: AI’s potential in medical ethics is considerable, but it must be wielded with care, understanding, and above all, human touch.

 

Paper title: Exploring the Potential Utility of AI Large Language Models for Medical Ethics: An Expert Panel Evaluation of GPT-4

Authors:

Michael Balas¹, Jordan J Wadden^2,3, Philip C Hébert^1,4, Eric Mathison⁵, Marika D Warren⁶, Victoria Seavilleklein⁷, Daniel Wyzynski⁸, Alison Callahan⁹, Sean A Crawford^1,10, Parnian Arjmand¹¹, Edsel B Ing^1,12

Affiliations:

¹Temerty Faculty of Medicine, University of Toronto, Toronto, ON, Canada

²Centre for Clinical Ethics, Unity Health Toronto, Toronto, ON, Canada

³Clinical Ethics, Scarborough Health Network, Toronto, ON, Canada

⁴Department of Family and Community Medicine, University of Toronto, Toronto, ON, Canada

⁵Department of Philosophy, University of Toronto Scarborough, Toronto, ON, Canada

⁶Department of Bioethics, Dalhousie University, Halifax, NS, Canada

⁷Clinical Ethics Service, Alberta Health Services, AB, Canada

⁸Office of Health Ethics, London Health Sciences Centre, Western University, London, ON, Canada

⁹Ethics Department, Ontario Shores Centre for Mental Health Sciences, Whitby, ON, Canada

¹⁰Division of Vascular Surgery, Department of Surgery, University Health Network, Toronto,

ON, Canada

¹¹Mississauga Retina Institute, Toronto, ON, Canada

¹²Department of Ophthalmology and Vision Sciences, University of Alberta, Edmonton, AB, Canada

Competing interests: None declared

Social media accounts of post author(s):

Twitter: @theMichaelBalas

Instagram: @michael.balas