Core Insights - Ant Group's CEO, Han Xinyi, emphasized the importance of specialized AI models in the healthcare sector, stating that general models cannot replace them in the short term [1][2] - The company aims to address key issues such as data quality, hallucination suppression, and medical ethics to enhance AI's role as a supportive tool for doctors rather than a replacement [2][3] Group 1: AI in Healthcare - Ant Group is focusing on the dual characteristics of "urgent need + high frequency" in healthcare, combining low-frequency medical actions with high-frequency health management to create a fertile ground for AI services [2] - The ultimate goal of AI in healthcare is to provide personalized, precise, and trustworthy recommendations akin to those of professional doctors, which general models will struggle to achieve for a considerable time [2][3] - Han Xinyi firmly stated that AI will not replace doctors in the foreseeable future but will serve as an assistant, helping specialists expand their capabilities and allowing primary care physicians to have better support [2][3] Group 2: Challenges in AI Healthcare Implementation - High-quality data is essential, with costs for data labeling and training potentially exceeding hundreds of dollars per data point, requiring involvement from senior medical experts to ensure quality [3] - Suppressing hallucinations in AI models is a significant challenge, where the goal is to reduce errors without compromising service capabilities, necessitating careful balancing [3] - Medical ethics presents a complex challenge, prompting Ant Group to establish a Medical Ethics Advisory Committee to explore regulations collaboratively with top experts in the field [3] Group 3: Market Position and Future Plans - The healthcare market is valued at trillions, but Ant Group is not rushing into commercialization; instead, it is prioritizing the accumulation of professional data, hallucination suppression, and ethical standards [3] - As of June 2023, Ant Group launched the AI Health Manager AQ, which has served over 140 million users, connected with more than 5,000 hospitals, and assisted nearly 1 million real doctors [5]
蚂蚁集团CEO韩歆毅外滩分享：AI医疗唯一的出路是人机结合