Related Experiment Video
Updated: Jan 15, 2026
08:06
Mixed Reality Assisted Radical Endoscopic Thyroidectomy
Published on: January 31, 2025
659
Expectations vs Reality of an Intraoperative Artificial Intelligence Intervention
Melissa Thornton1,2, Benjamin A Y Cher3,4,5, Cameron Macdonald6
1Department of Surgery, University of Texas Southwestern Medical Center, Dallas.
JAMA Surgery
|January 14, 2026
Summary
Bridging the gap between surgeon expectations and artificial intelligence (AI) program reality is key for successful implementation. This study identified key misalignments in AI surgical tools, offering lessons for future adoption.
Area of Science:
- Surgical Innovation
- Artificial Intelligence in Medicine
- Healthcare Technology Implementation
Background:
- Significant gaps between expectations and reality hinder the implementation of artificial intelligence (AI)-based healthcare programs.
- The Operating Room Black Box is a novel AI intervention designed to improve surgical outcomes.
- Assessing the real-world effects and surgeon expectations of AI interventions is crucial for successful adoption.
Purpose of the Study:
- To identify barriers and facilitators to implementing AI-based interventions for intra- and postoperative care.
- To assess the alignment between surgeon expectations and the actual performance of an AI intervention in surgical settings.
Main Methods:
- A qualitative, multisite study involving semistructured interviews with surgeons and implementation leaders at 3 academic centers.
- 30 surgeons and 17 implementation leaders were interviewed regarding their experiences with an AI intervention.
- Data were collected and analyzed between 2021 and 2024 to identify misalignments between expectations and deliverables.
Main Results:
- Four major themes of misalignment were identified: AI model training needs, difficulties in data access, limited predictive ability for postoperative complications, and few academic deliverables.
- Most surgeons had neutral (57%) or positive (37%) views of the AI technology.
- Participant demographics included surgeons aged 35-50 and older implementation leaders, with a notable proportion of female participants.
Conclusions:
- Successful implementation of AI interventions in surgery requires minimizing the gap between surgeon expectations and the technology's actual capabilities.
- Addressing these identified misalignments is critical for overcoming barriers to AI adoption in healthcare.
- This multisite assessment provides valuable lessons for optimizing the implementation of AI in surgical settings.
