1359 Clinical Use of Artificial Intelligence Among Irish Doctors

Abstract

With the growing use of artificial intelligence (AI) tools, the health service executive (HSE) has begun formally integrating AI into national digital and clinical governance structures. Despite this, real world patterns of AI use among doctors remain poorly characterised. We conducted a cross sectional, observational and descriptive study using an anonymous online questionnaire distributed to doctors working across Ireland. The survey captured demographic data, context of AI use and perceived influence on clinical decision making including benefits and concerns. Eighty-nine doctors consented and completed the survey. Seventy-four (83%) reported using AI during clinical care with fifty-two (58%) using it at least weekly. AI was used most commonly for rapid clarification of unfamiliar topics (71%), interpreting complex polypharmacy (60%) and checking medication interactions and doses (51%). Among those who use AI, chat GPT is most commonly used (82%). AI was reported to have a moderate to very strong influence formulating a differential diagnosis (49%), selecting investigations (45%) and in prescribing decisions (42%). Forty-eight (54%) reported identifying incorrect AI recommendations regularly. Information was most often verified by cross checking with colleagues (34%) or guidelines (27%). Sixty-four (72%) doctors expressed concern that AI may provide clinically unsafe information and seventy-five (84%) were unsure about how AI use fits within clinical governance. While clinicians recognise its efficiency and cognitive benefits, substantial concern remains regarding safety and regulation. These findings highlight the urgent need for clear national and hospital-level policies to support safe, regulated integration of AI into clinical practice.