AB060. SOH25_AB_019. Comparative analysis of perceptions on intraoperative artificial intelligence: a survey study amongst surgeons and medical students in Ireland

Abstract

Background: Artificial intelligence (AI) promises to revolutionise healthcare but has been previously characterised by cycles of ‘boom’ and ‘bust’. Alongside technological capability, realistic user expectations are essential for appropriate implementation. We surveyed surgeons, surgical trainees, and medical students in Ireland regarding their current perceptions. Methods: Electronic survey distributed through professional networks and social media with institutional ethical approval. Statistical and thematic analyses were performed to identify key perspectives. Results: Among 94 participants (63% medical students, 18% surgical trainees, 15% consultants), 62.7% “strongly agreed” that AI could enhance real-time decision-making during surgery. Most (90.5%) believed AI was already being surgically deployed to some extent although only 18% felt it appropriate ever to use for decision-making. While 53.2% were positive about AI’s potential to improve surgical outcomes, 72.3% reported no AI training in this context despite 86.2% (92.8% of consultants were “extremely/very interested”) expressing interest. The primary concerns with AI regarded accuracy and reliability (38.7%) and the lack of evidence of effectiveness (33.7%). Surgical trainees expressed greater concern about AI transparency (47% “extremely concerned”) compared to consultants (42.9% “slightly concerned”) and, along with students (both >50%) declared higher concern regarding liability issues versus consultants (64.3% of whom had “little to no concern”). Conclusions: Students and postgraduates in surgery in Ireland express optimism and high expectations for AI’s potential to improve surgery. However, concerns about reliability, evidence, and liability persist with clear caution regarding automated decision making and insight regarding need for education that may help align expectations realistically regarding AI evolution.

Ähnliche Arbeiten