Research On Social Sharing Of Artificial Intelligence Damage Risk

As a rapidly developing science and technology,artificial intelligence(AI)has brought many conveniences to human society,but it also inevitably brings some real or potential damage risks.Under the background of the unstoppable development of AI technology,it is the inevitable cost of scientific and technological progress that the damage consequences are shared by the benefit society.Due to the strong technical risk characteristics of AI damage,it brings great challenges to the current tort liability system.How to protect personal rights and promote the development of AI industry,in order to better enjoy the huge dividends brought by AI,is a major issue facing all countries in the world.For AI damage risk,the European Union,the United States and other developed economies have begun to explore a new liability framework,which not only breaks through the limitation of individual tort liability,but also jumps out of the barriers of tort liability system.In particular,the AI damage risk is dispersed in various social entities,and the concept of social relief is established.In order to make up for the deficiencies of the existing tort liability system in the field of artificial intelligence damage,and to enable producers and designers to focus on promoting technological progress under the premise of certain duty of care and product liability,it is necessary to introduce the socialized sharing mechanism of artificial intelligence damage risk in China.According to the research idea of "concept definition → problem investigation→ theoretical analysis → path selection → institutional construction",this paper closely focuses on the core theme of socialized risk sharing of AI damage,with the help of multi-disciplinary research tools and theoretical paradigms of institutional economics,sociology,law,etc.,through the analysis of practical cases of AI damage and the socialized risk sharing practice of foreign countries Combined with the relevant theoretical basis of social sharing,this paper attempts to build a comprehensive mechanism of social sharing of AI damage risk in China.This article uses the progressive development of the discussion.In order to clarify the concept category of artificial intelligence,through comprehensive consideration and comparative demonstration of the historical evolution and core elements of artificial intelligence,it is considered that the core elements of the concept of artificial intelligence should include: a system or device,thinking or acting rationally like human beings,can approach,achieve or even surpass the results of human behavior.Artificial intelligence is divided into strong artificial intelligence,weak artificial intelligence and super artificial intelligence according to the intensity of autonomous consciousness,high-risk artificial intelligence and low-risk artificial intelligence according to the possibility of causing damage,and autonomous artificial intelligence and human-computer mutual assistance artificial intelligence according to the degree of human participation In the evolution stage of intelligence to strong artificial intelligence,the damage risk of high-risk autonomous artificial intelligence.This paper also describes the connotation and main damage forms of AI technology risk,that is,the damage risk of AI is rooted in the essence of its technology risk,which can cause damage to the personal and property safety of uncertain personnel,and has a variety of risks such as ethics,alienation,responsibility,etc.It lists the damages in the application fields of automatic driving,surgical robot,intelligent investment advisor,etc.,for illustration The artificial intelligence damage risk has the technical characteristics different from the traditional damage risk,that is,it has the characteristics of system,potential,accumulation,pluralism,autonomy and complexity of causal relationship.Based on the technical risk characteristics of AI damage risk,the current tort liability system is still difficult to overcome its own limitations and unable to cope with the great challenges brought by AI damage even if it is constantly modernized.The deep-seated reason is that the current tort liability system is based on the value pursuit of "prevention and control of private to private interest infringement" and "victim centered",which adjusts the definite,limited,measurable,controllable and predictable private damage behavior,and the responsible party bears the independent and limited liability for compensation.The damage risk of artificial intelligence has the characteristics of publicity,universality,immeasurability,uncertainty and so on,which completely subverts the operating conditions of the tort liability system,and also causes the practical dilemma that the tort liability system is unable to deal with the damage of artificial intelligence.It is mainly reflected in the dilemma of subject definition,fault principle application,burden of proof distribution,product quality responsibility,causality determination and damage relief failure.It is difficult for the victims to get timely relief due to the inability to quickly determine the responsible party.Even if the responsible party is determined,it may face huge development and operation risks due to lack of compensation capacity.This leads to two related problems that need to be solved urgently: who bears the risk of AI damage can better reflect fairness and efficiency? How to carry out social sharing?Aristotle’s theory of distributive justice and the relief concept of "sharing-sharing" in modern society provide a solid legal theoretical basis for the socialized sharing of AI damage risk.The theory of distributive justice requires the state to ensure that everyone can enjoy certain material wealth.Social sharing reallocates social resources through insurance,fund and other forms,realizing the justice value of modern welfare state.The relief concept of "sharing-sharing" is the proper meaning of "sharing the benefits of scientific and Technological Development and sharing the costs of damage risk equally" in modern risk society.The analysis of law and economics provides technical support for social sharing.Both Calabresi’s accident cost allocation theory and Posner’s accident avoidance theory believe that accident risk allocation and accident prevention can be realized through social sharing.Through the derivation of "responsibility innovation" model,it shows that the imputation principle has an impact on the innovation and development of AI enterprises.The government can get out of the dilemma caused by the traditional imputation principle by introducing social sharing.Based on the theoretical analysis,it is necessary to clarify the orientation of the socialized sharing system in order to give full play to its functions of relieving victims,preventing damage and guiding industrial development.Through the empirical study of the first party insurance sharing,social insurance sharing,liability insurance sharing and compensation fund sharing,and further consideration and evaluation are carried out respectively,and the foreign experience can be drawn for reference.On the basis of full reference,the system of socialized sharing of artificial intelligence damage risk in China is constructed,and the basic principles of social sharing are legal principle,public welfare principle,relief principle,proportion principle and development principle.In the system architecture,it is suggested to set up a special artificial intelligence supervision agency and relief organization to conduct hierarchical supervision of artificial intelligence and the collection and use of damage compensation fund.In the way of socialized sharing,it is suggested that artificial intelligence liability insurance and artificial intelligence compensation fund be the main body,and the first party insurance and social insurance as the supplement.At the same time,the parallel source problem of each sharing mode should be dealt with.In order to effectively implement the social sharing system,it is necessary to make artificial intelligence compulsory registration,algorithm audit,black box security monitoring and defective product recall as the system supporting system.According to the overall framework design,this paper discusses the system construction of artificial intelligence liability insurance and artificial intelligence damage compensation fund.Artificial intelligence liability insurance system is the most important way and way of social sharing.This paper discusses the establishment basis,system structure and risk control.The purpose of AI liability insurance is to disperse the risks of artificial intelligence producers and designers,and compensate the victims in time,which is conducive to the enterprises to promote technological innovation and promote the development of artificial intelligence industry.Artificial intelligence liability insurance is divided into arbitrary liability insurance and compulsory liability insurance.Through comparative analysis,it is considered that the combination of the two models is more suitable for China’s national conditions.The compulsory liability insurance is carried out for "high risk" artificial intelligence,which is conducive to the prevention and control of damage risk and timely relief of the victims.We will carry out any liability insurance for "low risk" artificial intelligence and take the way of voluntary purchase.This paper constructs the liability insurance system from five aspects: the main body standard,the source of premium,the design of insurance type,the coverage and the differential floating rate.In the aspect of risk control,it mainly includes setting compensation limit and deductible amount,claim insurance in the selection period,introducing reinsurance sharing risk,carrying out risk sharing in the same industry,and monitoring the whole process risk.AI damage compensation fund system is another important pillar of social sharing system and a powerful supplement to AI liability insurance system.This paper constructs the system from five aspects: fund management organization,fund relief scope,application conditions and procedures,recovery mechanism and operation supervision.The purpose of setting up compensation fund is to reduce and share the responsibility risk of AI producers and designers,ensure the injured party to get compensation in time,and promote the development of AI industry.It is suggested that the compensation fund should be operated by the corporate body under the guidance of the AI Regulatory Committee.In terms of fund raising,it mainly includes enterprise payment,government appropriation,investment income,social donation and other sources to ensure the fund pool full.To clarify the scope,basis and time limit of compensation for victims’ damage.The compensation fund applies the principle of no fault compensation.The applicant who meets the scope of the application shall be compensated in time after the examination.The content of compensation includes personal injury and direct property damage.The compensation fund needs to be supervised by the AI Regulatory Commission and relevant internal and external departments to ensure its safe and smooth operation.