Current practices and perceptions of ChatGPT in gynecologic oncology: results from a cross-sectional questionnaire (TRSGO-AI-001)

Abstract

Background We aimed to assess current practices, perceptions, and perceived effectiveness of ChatGPT among gynecologic oncology professionals in academic and clinical settings. Methods A 23-item, international online survey was conducted between July and September 2025. Gynecologic oncology professionals were recruited via social media and snowball sampling. We evaluated demographics, awareness, utilization, and perceptions of ChatGPT-3.5 across different domains. Results Our respondents( n = 111) were predominantly aged 36–45 years(55.0%), from Türkiye (90.1%), and affiliated with academic institutions(87.4%). Nearly all respondents(99.1%) were aware of ChatGPT, and 66.7% reported professional use, primarily for literature summarization(52.3%), teaching(42.3%), and academic writing(36.9%). Direct clinical use remained somewhat limited(17.1%), with minimal engagement in patient communication(7.2%). Misinformation was the most commonly reported concern(66.7%). Perceived effectiveness was highest in academic writing, literature summarization, and teaching, whereas patient communication, clinical decision support, and exam preparation were rated somewhat low. Overall, 60.3% indicated that ChatGPT contributed meaningfully to their work, 45.9% supported its integration into gynecologic oncology education, and 58.5% would recommend it to colleagues. Participants < 45 years and those with < 5 years of clinical experience reported significantly more frequent ChatGPT use in their professional activities(p-value = 0.026). Those with greater clinical experience were more likely to use ChatGPT for writing purposes(p-value = 0.039), whereas those < 45 years used it more often for clinical decision-making(p-value = 0.015). Conclusions ChatGPT is widely recognized and used for academic and educational tasks in gynecologic oncology, yet its clinical applications remain limited. Future research should optimize large language models for clinical use, evaluate comparative outcomes across diverse models, and investigate their integration into multidisciplinary care.