Ethical considerations for deploying AI in surgery

Editor's Note

A balanced approach to leveraging AI in surgical decision making should be taken to weigh the benefits of increased efficiency and potential accuracy of using AI against the principles of medical ethics such as patient autonomy, beneficence, nonmaleficence, and justice, authors recently advised in the November issue of the Journal of Surgical Research.

As the medical field learns and begins to integrate AI into surgical care, the authors suggested bias, inequality, transparency and patient safety must remain top of mind. To assess a present level of engagement of AI in the surgical community, the authors surveyed attendees during a symposium on AI at the 2024 Academic Surgical Congress in Las Vegas, NV. The audience was asked two questions:

*“Have you used AI in any form recently?”

*“How many of you are using it in your practice currently?”

Of 52 respondents to the first question, the authors reported 92% had used any form of AI recently for any purpose. Of 50 participants who responded to the second question, 56% reported they haven’t used AI in their clinical practice. A smaller percentage of respondents (12 out of 50) were using AI for teaching in the operating room, for clinical-decision support, or for clinical administrative support. Around one third of surveyed attendees reported researching the uses of AI, pointing to a particular interest in exploring and expanding AI's role in clinical settings.

The authors concluded that while AI is being used in health care, not all health care providers are using this technology routinely. During their broader panel discussion where the survey was conducted, several other themes were discussed that are relevant to future implications for AI in clinical and surgical care, including:

*Fairness and equity in access

*Informed consent

*Data stewardship

*Accountability

The authors further suggested, per the report, that surgeons must take ownership and be the final arbiter of AI-assisted technology in the care of surgical patients. In these efforts, they suggested that work toward a single standard for AI quality reporting should consider emerging frameworks in the European Union and the United States that mandate disclosures on data governance, risk management, and performance for healthcare AI.