China tests chatbot for surgeons

AI assistants such as ChatGPT and Microsoft Copilot are already used for daily tasks. Yet, what if more important roles, such as healthcare, are created? That is what China’s Center for Artificial Intelligence and Robotics aims to develop. The country is testing its healthcare-focused AI program at seven different hospitals in Beijing. 

Arguably, the most interesting aspect of the AI project is it will make a chatbot for surgeons. One day, a doctor may speak with an AI program to bring up important diagnostic data during an operation. Even better, your physician may ask their bot to guide their medical decisions!

How will this chatbot for surgeons work?

Yahoo Finance reported a Hong Kong-based agency under the Chinese Academy of Sciences introduced an AI model based on Meta’s LLAMA 2.0 AI model. 

Executive director Liu Hongbin said he and his colleagues trained the model with papers, medical journals, and manuals so that it could become a surgery consultant. 

Moreover, they powered training with 100 graphics processor units, 50 Nvidia A100 chips, and 50 Huawei Technologies’ Ascend 910B chips. 

Eventually, they developed the CARES Copilot 1.0. Interesting Engineering says the bot will answer various medical questions. 

The program will cite sources based on more than a million academic records. Also, CARES will process diagnostic data, such as audio, text, images, CT scans, MRIs, and ultrasounds.

More importantly, scientists say they want the medical chatbot to serve more active roles. Specifically, they want the bot to provide medical advice to physicians when treating patients. 

CARES will become an assistant that warns doctors about performing risky procedures that may cause harm. 

READ: Robots are transforming medical education

Feng Ming, chief physician at Peking Union Medical College Hospital’s neurosurgery department, commented on the AI bot:

“There are obstacles, including restricted computing power due to the banned access to NVIDIA’s advanced chips.”

“However, we can develop a vertical model with our own characteristics with more high-quality data from top hospitals in the mainland, which is not available for OpenAI and many domestic private companies.”