Abstract
New artificial intelligence (AI)/machine learning (ML) technology offers great potential to assist surgeons with real-time intra-operative decision-making. While, AI/ML-driven analysis tools for surgeons currently focus primarily on technical assistance and postoperative insights, AI/ML cognitive support in surgery can improve their usefulness. However, AI/ML models usually conceal their underlying algorithmic reasoning process. As a result, such "black box" AI/ML models have important clinical and legal implications for patient safety and surgeon liability. This article provides an overview of surgeons' current practice and the potential for AI enhancement in surgical decision-making. It suggests a path toward a safe and effective integration of black-box AI/ML models into the operating room. We posit that future surgeons who rely on AI for cognitive assistance do not necessarily need to fully understand, interpret, and explain the algorithmic basis of an AI's real-time recommendation in the midst of surgery, but rather, they need to know that these tools work as promised. Assuming new black-box AI/ML models demonstrate clear benefits for surgical patients, their use will likely be incorporated into the legal standard of care and affect the liability landscape for surgeons.
Original language | English |
---|---|
Pages (from-to) | 2823-2826 |
Number of pages | 4 |
Journal | International Journal of Surgery |
Volume | 111 |
Issue number | 4 |
Early online date | 24 Feb 2025 |
DOIs | |
Publication status | Published - 1 Apr 2025 |