How will generative artificial intelligence (AI) drive transformative paradigm shifts in the areas of clinical care, practice management, surgical education, and research?
In the Opening Day Thematic Session, “Generative AI Tools for Surgery: Will AI Change My Practice?,” October 20 at 11:30 am in Room 303-304 South, presenters will discuss applications of this technology—from ambient AI for automated documentation to analysis of surgical video and records for enhanced prediction, assessment, and feedback data.
“I hope that attendees will get an accurate picture of the current landscape of generative AI technologies,” said moderator Daniel A. Hashimoto, MD, MS. “It is our goal for attendees to learn how existing generative AI tools can be used and how they should not be used in situations where they have not been adequately vetted. We want this session to offer an ‘under the hood’ view of these algorithms and provide practical suggestions on their applications in surgical practice.”
Generative AI relies on large datasets while healthcare-specific AI models typically tailor those general models to fit distributions in medical datasets—which are not only smaller, but also more limited in scope.
“A lot of AI drills down to the quality of the data you are using to build these models,” explained co-moderator Caroline Park, MD, MPH, FACS. “Trash in, trash out. Be wary of where the data are coming from and if they represent your population. The more heterogeneous the data, the better the algorithm.”
Panelists representing several key facets of AI implementation in medicine will include leaders in the health informatics space with expertise in generative AI, biomedical informatics, and regulatory considerations for healthcare-related software. This session also will feature a robust Q&A in which panelists will address audience questions regarding how this technology may impact surgeons' and patients' lives.
“We have a great lineup of speakers,” said Dr. Hashimoto. “Dr. Genevieve Melton-Meaux, president of the American Medical Informatics Association, will provide her expertise as both a colorectal surgeon and leading health informaticist. Dr. Gabriel Brat, a trauma surgeon, associate editor of NEJM AI, and Vice-Chair of the AI Subcommittee of the ACS Health IT Committee, will offer his perspective on how surgeons could benefit from AI analytics. And Dr. Tyler Loftus, Chair of the ACS Health IT Committee and a trauma surgeon, has extensive experience in AI surgical risk calculators and will provide on-the-ground views of how these calculators are changing clinical care today.”
Industry experts suggest a collaborative approach is paramount in order to successfully guide generative AI implementation into surgical practice.
“This space is all about collaboration,” said Dr. Park. “The best partnerships include stakeholders from industry, regulation, administrative, and leaders who work in the patient- and provider-facing spaces. Creating and executing plans in a bubble is helpful only to a small group of people. We are excited to be at the tip of the spear in encouraging collaborations outside the ACS.”
The session also will identify relevant ethical and regulatory considerations related to integrating generative AI into clinical settings.
“Attendees will learn about potential areas of bias that may arise from generative AI models that can lead to inaccurate analyses or outputs from generative models,” said Dr. Hashimoto. “We will discuss regulatory considerations around when a model might be considered safe enough for clinical deployment or what potential pitfalls may arise once models are deployed. Importantly, the panel also will discuss potential impacts on malpractice and scope of practice.”
