Bridging the gap: Usage patterns and training needs of generative artificial intelligence among oncologists in Argentina.

e13714 Background: Generative Artificial Intelligence (GenAI) is being rapidly integrated into clinical oncology practice. Recent publications highlight its potential benefits and risks in clinical decision-making. Understanding stakeholders' perceptions in low- and middle-income countries (LMIC) is critical for developing safe and effective tools. Methods: A cross-sectional survey was conducted among volunteer participants attending the Argentinian Association of Clinical Oncology (AAOC) annual meeting 2024. The objective was assessing perceptions and usage of GenAI tools among attending physicians. The survey was developed following a literature review and feedback from the AAOC Digital Health & Artificial Intelligence Commission. Descriptive statistical methods were used to summarize the results. Results: Among the 164 respondents, 51% were male, 49% female; the median age was 43 years (range 26-65). Most of the population (91%) were medical oncologists. Participants included 30% with less than 10 years of experience, 38% with 10–20 years, and 32% with over 20 years. The majority (61%) had used GenAI tools in the past month, with 21% reported daily use. Eighty-two percent had not received formal training on the use of GenAI in healthcare, yet 94% believed it could be useful. The highest usage was observed among younger participants and those with fewer years of medical experience (p = 0.03). Most of the physicians (79%) believed that GenAI tools have the potential to improve oncology practice, and 52% expressed confidence in the diagnoses provided by GenAI. Only 4% expressed concern about these tools directly guiding patient care, but 55% were concerned about patients following GenAI advice instead of receiving medical guidance. 84% believed these tools should be regulated, while 71% expressed concern about legal liability for errors made by AI. Conclusions: This cross-sectional survey highlights the growing adoption of GenAI tools among physicians in Argentina, particularly among those who are younger and less experienced. These findings suggest that while GenAI is trusted for diagnosis, concerns regarding patient misuse and associated legal issues highlight the need for specific training and robust regulatory frameworks. Developing knowledge and resources tailored to the local needs of LMICs is essential for improving access and the effective applying these new technologies.