信任与怀疑之间，我们如何与AI医生共存？

Core Viewpoint - The rise of AI in healthcare presents both opportunities and risks, with users increasingly relying on AI applications for medical advice, but also facing potential dangers from inaccurate recommendations [1][3][5]. Group 1: AI in Healthcare Applications - Numerous AI diagnostic applications have emerged, with 101 medical AI models registered in China, providing easy access to healthcare advice without the need for professional equipment or appointments [3][14]. - AI is particularly effective in diagnosing common ailments like colds and stomach pain, helping to alleviate healthcare anxiety for users, especially those in areas with limited medical resources [3][14]. Group 2: Risks and Challenges - There have been alarming cases where users suffered serious health issues due to following AI-generated advice, such as a user developing bromine poisoning from incorrect dietary recommendations [5][8]. - AI's limitations in accurately diagnosing complex conditions can lead to misdiagnosis, as it relies on user-submitted information that may be incomplete or inaccurate [9][11]. Group 3: AI's Role in Medical Decision-Making - AI lacks the ability to perform comprehensive assessments that human doctors do, such as physical examinations and nuanced understanding of patient history, which are critical for accurate diagnosis [9][11]. - The technology is primarily trained on common diseases and may overlook rare but serious conditions, potentially delaying critical treatment [11][13]. Group 4: Industry Response and Future Directions - The industry is exploring solutions to improve AI's reliability, such as using high-confidence databases for training and shifting AI's role to support rather than replace human doctors [14][15]. - Current AI applications should focus on low-risk scenarios like medication reminders and health monitoring, while avoiding involvement in complex or urgent medical situations [17][19].