Artificial intelligence and chatbots in general surgery: a survey among surgeons in Germany, Austria and Switzerland

Background: The integration of artificial intelligence (AI) and large language models (LLMs) into surgical practice is increasingly being explored, but real-world usage patterns and barriers remain insufficiently characterized. This multinational survey assessed self-reported AI/LLM use, perceived benefits, and requirements for routine implementation among general surgeons in German-speaking countries. Methods: Between June and September 2025, a 16-item online survey was conducted among general surgeons at university hospitals in Germany, Austria, and Switzerland. Results: Of 3831 invited surgeons, 323 complete responses were analyzed (response rate 8.7%). Self-reported AI use was frequent: 58.5% reported occasional and 28.2% regular use. The most frequent applications were speech recognition (65.3%) and chatbots (62.8%). Anticipated benefits focused on documentation simplification (94.4%), reduced administrative time (84.2%) as well as burden (83.0%), and improved diagnostic accuracy (70.6%). ChatGPT was the leading chatbot (89.8%), chatbot use was rated helpful (69.6%), and the most common use case was scientific writing (51.4%). Key barriers to routine AI adoption were insufficient integration into existing systems (77.1%), legal/data-protection uncertainty (65.9%), and lack of validated applications (59.1%). The most important requirements were system reliability (76.2%), a clear legal framework (72.1%), improved technical infrastructure (68.4%), and transparency (58.5%). Most respondents expected AI to improve surgical care quality (82.4%) and endorsed structured AI training (85.1%). Conclusion: The use of AI and LLM-based chatbots was commonly reported among general surgeons in German-speaking countries, primarily for low-threshold, efficiency-oriented tasks such as speech recognition, documentation, scientific writing, and other text-based productivity. Broader integration into surgical workflow will require interoperable implementation, validated clinical-grade applications, legal clarity, and structured education.