On the morning of January 22, at the "Baichuan-M3 Plus Model Release Media Briefing," Baichuan Intelligent, following the open-sourcing of its new-generation medical large language model Baichuan-M3 just nine days prior, unveiled the Baichuan-M3 Plus model. The company announced that in the Hallucination Rate evaluation, the M3 Plus model's hallucination rate has been further reduced to 2.6%, while API call costs have been cut by 70%.

During discussions with media outlets including Sina Technology, Baichuan Intelligent's founder and CEO Wang Xiaochuan commented on the impact of large AI models on the healthcare industry. He stated, "I was recently asked by the media about my views on Zhang Wenhong's discussion regarding restricting doctors' use of AI. I've since re-watched the video. While I may not fully grasp his original intent, I have given deep thought to his public statements."

Wang Xiaochuan expressed directly, "I believe doctors universally acknowledge the principle of patient interests first. AI is advancing at a remarkable pace. In certain scenarios, the combination of AI and a doctor has demonstrably surpassed the capabilities of a doctor working alone, which aligns with the principles of medical science. Restricting the use of AI due to concerns about hindering 'doctor development' might consequently restrict the medical measures that are most beneficial to patients. If AI can provide substantive assistance to current patients, its use should not be rejected. In other words, the development of doctors should not come at the cost of current patients."

From Wang Xiaochuan's perspective, the issue is not AI itself, but rather the exploration of how to use it effectively; determining the optimal application of AI is a key challenge. If the concern is the potential degradation of doctors' skills, perhaps the approach should be reconsidered: when junior doctors begin their practice, instead of requiring them to correct AI, the model could be used to provide reminders on their clinical reasoning and to verify diagnostic outcomes. This approach could more effectively reduce missed diagnoses, misdiagnoses, and associated risks, thereby benefiting both patients and facilitating the growth of the doctors.