The 2023 World Artificial Intelligence Conference, recently held in Shanghai, set a new record for the sixth consecutive conference with over 400 participating companies and a 50,000-square-meter main exhibition area. "AI + Rehabilitation Robots," "AI + Drug Development," "AI + Image Scanning"... from inside the exhibition hall to outside, from laboratories to clinics, artificial intelligence is accelerating its integration with medical innovation.
An investigation by reporters from the Economic Information Daily found that as artificial intelligence moves from perception to cognition, from recognition to generation, and from specialized to general applications, it will be deeply expanded in the medical field in the future, creating more "super" capabilities and broadening the scope of life and the depth of health.
A constant stream of healthy and effective products are emerging.
After the release of the domestic sci-fi film "The Wandering Earth 2", the exoskeleton robot that gives humans "superpowers" has repeatedly made headlines. In real life, an exoskeleton robot for rehabilitation has been "implemented" in many medical institutions across the country.
This robot, featuring an integrated ergonomic design, helps patients with various types of lower limb motor disorders achieve early training in standing and walking. It integrates actuators with motors and reducers, intelligent algorithms that ensure stable coordination between actuators, and a mechanical structure that realistically simulates the floating of the pelvis during walking. It provides force assistance when patients attempt voluntary movement and can adjust various gait parameters such as stride height, stride length, and stride speed according to different patient conditions, activating the patient's neural regulation and muscle recruitment to improve rehabilitation outcomes.
Beyond its visible "hard armor," artificial intelligence also relies on its invisible and intangible algorithms and data as core competitive advantages. XtalPi, the winner of the SAIL Award, the highest honor at the 2023 World Artificial Intelligence Conference, showcased its independently developed "Intelligent Automated Drug Discovery Platform": drug molecules designed by AI algorithm models will be manufactured by automated robots; standardized, recorded, traceable process data and results will continue to "feed" the AI model for learning and iteration, helping more new drug development go from "0 to 1".
Zhang Peiyu, Chief Scientific Officer of XtalPi, stated that in the past, developing a new drug typically required 10 years and cost $1 billion, with a success rate of less than 10%. By leveraging artificial intelligence technology to empower new drug development, costs can be reduced, efficiency improved, and the success rate increased. For example, in the development of a small-molecule oral drug, the XtalPi team collaborated with Pfizer to identify the drug's optimal crystal form in just six weeks, accelerating its market launch. Currently, XtalPi has participated in the innovative drug pipelines of numerous domestic and international companies, encompassing over 180 drugs, including small molecules, antibodies, and peptides.
Recently, InSilicon Technologies announced that its anti-fibrotic small molecule drug candidate has completed the dosing of the first patients in its Phase II clinical trial. This means that the candidate drug, which uses generative artificial intelligence to discover novel targets and design molecules, has advanced to the Phase II clinical trial validation stage. "AI-driven drug development has shown great potential in shortening the R&D cycle and reducing R&D investment," said Ren Feng, co-CEO and chief scientific officer of InSilicon Technologies.
New technologies empower new healthcare
There is a "special" department at Zhongshan Hospital affiliated with Fudan University. Based on a disease-specific database and expert knowledge graph, it combines computer graphics, graphics rendering, deep learning, speech synthesis and other technologies to create a "digital avatar" for outpatient doctors and integrate it into smart devices to provide virtual consultation services. This is the digital twin AI department.
"In the future, based on generative artificial intelligence technology and a rich medical knowledge database, the next generation of 'digital doctors' can play a highly efficient role in medical diagnosis, medical popularization, and patient education," said Gu Jianying, Party Secretary of Zhongshan Hospital. She added that, in order to build a national medical center and address pain points in healthcare services, Zhongshan Hospital is accelerating its embrace of new technologies and empowering new healthcare, deeply cultivating the construction of a smart hospital from multiple dimensions, including smart triage, smart outpatient clinics, smart wards, digital operating rooms, and imaging cloud.
"With the continuous emergence of large-scale models, generative artificial intelligence, and new business models, the future development of the health industry is entering a critical period of opportunity," said Zhang Ying, Deputy Director of the Shanghai Municipal Commission of Economy and Informatization, at the Health Summit Forum of the 2023 World Artificial Intelligence Conference. She added that Shanghai will actively promote the deep integration of artificial intelligence and healthcare, develop AI algorithms and large-scale models applicable to life science problems, support intelligent analysis and decision-making in different scenarios and tasks, promote innovative applications in fields such as medical imaging and drug development, and improve the level of intelligent healthcare.
In the future, large-scale AI models will become a transformative force driving productivity improvements. Min Dong, Deputy Director of the Cloud Computing and Big Data Institute of the China Academy of Information and Communications Technology (CAICT), stated at the health summit forum that large-scale models have already been explored to varying degrees in fields such as medical services, operations management, traditional medicine, scientific research, and drug supply. Going forward, CAICT will support relevant ministries in carrying out work related to large-scale AI models for healthcare, constructing standard frameworks, organizing collaborative research among relevant units, conducting application pilots, and identifying outstanding cases.
Application and regulation should proceed in tandem.
Industry insiders believe that while artificial intelligence technology brings new applications to the medical field, it will inevitably bring new challenges to medical quality and safety. Therefore, while helping the industry get on the "fast track", it is necessary to fasten the "seat belt".
Min Dong stated that current limitations in artificial intelligence technology may lead to content distortion or bias, while insufficient computing power may result in lower efficiency and higher costs. Therefore, a parallel approach of promotion and regulation should be adopted to create a safe, compliant, controllable, and reliable environment for the development of large-scale models. For example, in terms of algorithms, it is recommended to promote the construction of algorithm platforms and the deployment of computing infrastructure, foster industry consensus, and support the high-quality development of large-scale health and medical models.
In addition, regarding data security, which is of utmost concern to the public, besides building standardized and regulated disease specialty databases to ensure the quality of training data and privacy protection, it is also necessary to promote the compliance of AI medical applications, such as establishing a generated content detection mechanism to prevent algorithmic bias from affecting the fairness of medical services.
"It is foreseeable that the application prospects of artificial intelligence combined with medicine are broad in the future. However, it is also necessary to recognize that medicine is different from other industries, especially in terms of professionalism and precision. New technologies enabling new applications cannot be achieved without the collaboration and support of various parties, including clinicians, enterprises, and the legal system. Efforts need to be made to strengthen the foundational work and improve the market system for medical data elements, so that smart healthcare services can be readily accessed and better serve the health of the people," said Gu Jianying. □ Reported by Gong Wen and Du Kang, Shanghai
