标签： 阿拉上海后花园论坛

Generative AI(Generative AI) has set off a whirlwind in various fields, and it is considered to be a promising medical application. The industry admits that compared with the standard answer of judgmental AI, generative AI needs to overcome regulatory issues, and may be a possible application scenario in the short term, such as patient care, medical record/report generation and writing, and initial screening of telemedicine. Whether new drugs can be clinically developed in the future will have a more breakthrough development possibility.

Wang Dihao, director of Changjia Intelligent R&D, pointed out that the development of generative AI in the medical field is gradually moving forward. At present, it has been touched by people in the fields of text photos, pictures, sounds, etc. What is relatively new recently is the breakthrough in details, such as more simulated fingers, etc. However, compared with the standard answer of judgmental AI, which is still the mainstream of medical applications, generative AI does not produce an absolute answer of "yes or no", and it is difficult to be supervised in medical care.

However, Wang Dihao also believes that even if the regulatory issues still need to be broken, the generative AI will still have significance for the medical industry, including patient care, shortening the writing time of medical records and reports, and initial screening of telemedicine, which can help medical staff deal with related processes. It is more accurate and smooth, and it is still a scene that can be applied in medical care. However, he also admitted that there are many plans in the world, but no mature products have been seen yet.

The venture capital industry also believes that although the medical data startup Syntegra completed the fundraising in the first quarter of last year, and the Taiwan Province digital pathology and medical picture startup Yunxiang also completed the A+ round fundraising at the end of October last year, the market is still paying attention to the business model that similar startups can make profits in the future.

The industry believes that the most widely used AI in the field of intelligent medical care is still replacing large-scale repetitive administrative work or diagnosis. Although the COVID-19 epidemic has accelerated the development of telemedicine, this part still involves the integration of hospital organizational culture, governments of various countries and medical insurance payment norms, and there are still many places to be broken.

The international technology giants are looking forward to the progress of generative AI in the clinical development of new drugs. DeepMind, the most well-known artificial intelligence company under Alphabet, the parent company of Google, quickly predicts and generates a reasonable protein serial through neural network, and plans to sell or lease its AI technology successfully simulated in protein to new drug companies in the future to accelerate the clinical development of new drugs. It is planned to assist the new drug development company to develop the combination of drugs and protein serial. However, after the new drug development company was established independently, the market is still paying attention to its subsequent specific development.

Wang Dihao also said that although short-term generative AI has more room for development in non-medical core or medical materials that need to be approved, there are also opportunities for combining long-term judgmental AI with generative AI. For example, Changjia Intelligent develops the technology of judging whether there is bleeding or brain symptoms from brain computed tomography pictures, which can automatically output pictures into radiology reports by combining generative AI, greatly improving the efficiency of doctors’ interpretation and assisting in improving the diagnostic rate.