Abstract
Purpose
To describe perceptions and use regarding generative artificial intelligence (GAI) among oncology professionals across two surveys at the Argentine Association of Clinical Oncology (AAOC) meetings (2024-2025).
Methods
Two independent cross-sectional surveys were administered on-site using an expert-consensus instrument, assessing demographics, GAI use, frequency, trust, perceived impact, training needs, and barriers. Categorical data are reported as frequencies and percentages with Wilson 95% CIs; between-year comparisons used Fisher's exact test (exploratory).
Results
A total of 268 valid responses were analyzed (2024: n = 164; 2025: n = 104); 48.1% of respondents were male, 35.4% were age 36-45 years, and 40.7% had 10-20 years of experience. GAI use in the previous month was 69.0% (95% CI, 63.3 to 74.3) overall, 61.0% (95% CI, 53.3 to 68.1) in 2024 and 80.8% (95% CI, 72.2 to 87.2) in 2025. Near-daily use rose from 12.2% (95% CI, 8.0 to 18.1) to 33.7% (95% CI, 25.3 to 43.2) and was more frequent among younger participants (
P
= .048;
P
= .009). Main applications include administrative tasks (54.9%), scientific retrieval (41.8%), clinical consultation (15.3%), and patient information (7.1%). Most (79.1%, 95% CI, 73.8 to 83.5) had no formal GAI training; 92.9% (95% CI, 89.2 to 95.4) considered it necessary and 82.5% (95% CI, 77.5 to 86.5) believed GAI can improve cancer care. Trust in AI recommendations was higher in 2025 (67.3%, 95% CI, 57.8 to 75.6) than in 2024 (51.8%, 95% CI, 44.2 to 59.3). Privacy concerns (87.3%, 95% CI, 82.8 to 90.8), legal liability (70.1%), and support for regulation (84.0%, 95% CI, 79.1 to 87.9) were widespread.
Conclusion
GAI use was high and increasing, occurring without institutional support or formal training. These findings support educational program development and context-appropriate guidance for responsible adoption. To our knowledge, this is the first study characterizing GAI perception and use among oncology professionals in Argentina.