Research On The Regulations Of Automated Medical Diagnosis Algorithm Explanation

Because of the advantages of high efficiency,cost saving and resource balancing,automatic medical diagnosis has become the key area of the development of artificial intelligence.In the global novel coronavirus pneumonia battle in 2020,the advantages of automated medical diagnosis were revealed.The algorithm of automated medical diagnosis is gradually developed from the pre-set program of the research personnel to the deep learning of the self.But in the process of its development,there are many problems,such as misdiagnosis,algorithm non transparency,algorithm discrimination and so on,which make algorithm explanation necessary.When the relative of automatic decision-making requests algorithm explanation,it will lead to the leakage of medical sensitive data,infringement of trade secrets and other issues.How to define the algorithm explanation in the field of automated medical diagnosis is an issue that needs to be studied urgently.China has a certain legislative basis for the establishment of explanation regulations of automated medical diagnosis algorithm.Relevant legal provisions can be obtained from the constitution of the people's Republic of China,general principles of the civil law of the people's Republic of China,and network security law.Normative documents in the industry also give guidance,requiring the disclosure of algorithm information and strengthening the management of medical data.However,these regulations only mention the algorithm explanation of automated medical diagnosis,and there are still problems of lack of pertinence and coverage,which makes legal relief difficult.The EU adopts unified legislation and standards,and Member States and regions can also formulate their own targeted industry protection regulations under the EU legislation.GDPR plays a role in easing the contradiction between algorithm explanation and developers in the era of algorithm decision-making.Under the overall guidance of the EU,we should build the overall framework of algorithm governance,strengthen supervision,and supervise algorithm developers to comply with data protection,so as to promote the development and utilization of medical data.In the United States,the AI program is more focused on industry self-discipline,and there are mature system specifications for automatic image diagnosis algorithm.In addition,the United States adopts high-intensity privacy protection of medical data,and increases punishment.Constantly build and improve the supervision framework of artificial intelligence medical equipment,and summarize the experience and risk to ensure the safety and fairness of the implementation of the algorithm.Under the premise of maintaining the leading position of automated medical diagnosis,China should learn from the high standard and strict requirements of the explanation regulations of the European Union automated medical diagnosis algorithm,and learn from the flexibility of the US regulations.We are committed to solving the problems of scattered rules,weak applicability and strength.For the highly sensitive data leakage involved in the process of algorithm explanation,aunified and targeted regulation and department should be established.And grasp the "degree" of algorithm explanation,do not over disclose algorithm information,and seek a balance between algorithm explanation and protection of trade secrets and other sensitive data,so as to achieve information symmetry between algorithm counterparts,users and developers.