Research On Artificial Intelligence Tort Liability

In order to ascertain whether the current tort liability law can be effectively applied to high-tech AI infringement incidents,this paper uses historical research methods and Chinese-foreign comparative research methods to put forward China's responsibility rules for AI infringement incidents on the basis of historical experience and the legislative experience of other countries.Efforts will be made to promote the development of science and technology while also protect the rights and interests of victims.It is not yet to know whether the strong AI that is comparable to the level of human intelligence will come or not,and the super AI will not and shouldn't come due to people's fear and precaution for it's out of control.For now,the weak AI which lacks human rational thinking is still unable to reach the level of human universal intelligence.The loss of self-awareness and independent property makes it lose its capacity for civil conduct and civil liability.Giving the AI civil subject qualification is not in line with the people-oriented legal purpose and lacks the practical significance,and it is more likely to become the algorithmic dictator of R&D personnel,therefore,the AI is not enough to obtain civil subject qualification at present.When there is an AI infringement incident,AI itself cannot become the subject of infringement liability,but should be held responsible by the relevant subject behind AI according to the infringement reason.When AI infringes due to product defects,product responsibility should be borne by producers and sellers.At the same time,when identifying product design defects and warning defects,corresponding adjustments shall be made according to the particularity of AI,and it shall be made clear that the producer can invoke "development risk defense".When the owner or user causes the infringement of AI due to fault,they should be held accountable according to the principle of fault liability.When the autonomous behavior of AI causes infringement,it should apply fair liability to producers,sellers,owners or users to make up for the losses of victims and achieve the purpose of social justice.When the AI with a certain degree of autonomy and risk causes an infringement while cannot find out the reason of it,it's a more appropriate choice to apply high-risk liability regulation.The supporting measures taken for the particularity of AI mainly include the introduction of AI mandatory liability insurance and AI compensation liability fund to diversify the legal risks of the liability subject;In order to simplify the investigation of the cause of an accident,producers may be required to install "black box" to record data for AI.Consider developing ethical guidelines for AI R&D personnel to prevent or reduce AI infringement incidents at the source.Considering that AI may run out of control,producers should install a“one-button fuse” mechanism for them and take recall measures for AI that is out of control.