Dr. Courtney Balentine: "The Operating Room BlackBox: Implementation and Effectiveness of an Intraoperative AI Intervention"

SDSC’s Data Science Roundtable with Dr. Courtney Balentine – board-certified endocrine surgeon and associate professor at the University of Wisconsin School of Medicine and Public Health – explored the promise and practical challenges of using AI in real-time during surgery. His talk focused on his recent research on an “operating room black box” system that captures audio, video, and clinical data during procedures in an effort to improve quality and safety.¹

Dr. Balentine highlighted the growing tensions between expectations of AI as a transformative, near-autonomous tool and the reality of its current limitations. While the black box can generate rich datasets and support performance review, research shows it requires significant customization, training, and human interpretation before it can deliver meaningful insights.

Evidence suggests the technology can enhance surgical team feedback, identify workflow inefficiencies, and even correlate process quality with patient outcomes. However, implementation challenges are substantial. These include delays in accessing processed data, difficulty linking intraoperative events to postoperative outcomes, limited academic outputs, and user frustration when expectations are unmet.

Dr. Balentine noted the gap between the anticipated and actual value of an intraoperative AI intervention. Many clinicians expected real-time, automated insights, but instead encountered slow turnaround times and complex data handling. Successful adoption depended heavily on implementation strategies, including clear communication, realistic expectation-setting, and adapting workflows.

Dr. Balentine emphasized that early value may lie less in predicting complications and more in improving teamwork, communication, and training. Ongoing work aims to expand stakeholder engagement, explore legal and patient perspectives, and develop AI-driven coaching and simulation tools. 

‍

^{1. Thornton M, Cher BA, Macdonald C, Baker JG, Marten EL, Mai D, et al. Expectations vs reality of an intraoperative artificial intelligence intervention. JAMA Surgery. 2026 Mar 1;161(3):234. doi:10.1001/jamasurg.2025.6029}