The Civil Liability Caused By Artificial Intelligence

At present,the rapid development of science and technology has brought us into a new era of Artificial Intelligence.Through algorithms,artificial neural networks and other tools,artificial intelligence is autonomous,unpredictable and uncontrollable.The rapid development of artificial intelligence has also brought a series of political,economic,ethical and legal issues.This essay focuses on the issue of who should assume responsibility when an autonomous AI entity causes damage and whether the existing civil legal framework is sufficient to allocate responsibility.This essay studies the application of traditional types of civil liability in two different ways.The first involves integrating artificial intelligence into the traditional categories of products,animals and children but encountering difficulties: the last two are not enough to include AI,while the first one faces the inherent problems of artificial intelligence autonomy.However,AI is developing from simple automation to new entities with a high degree of autonomy,completing a qualitative shift from automation to autonomy.They are unique compared to the software embodied into the body.However,the intelligence of current AI systems is not enough to qualify them independent legal status.Therefore,although they are unique,they still belong to the category of products.The second method uses a different set of tools to deal with the legal implications of artificial intelligence,such as the contract used in order to allocate risks.On the other hand,insurance can address the different issues related to the compensation of victims.However,the cost of such insurance affects users,which may result in hindering the development of new technologies.When the AI entity causes damage,the compensation fund is a good solution to the victims' compensation.The right of information plays an important role in this area,not only during sales,but also at all stages of consumer engagement,including advertising.Among the two different approaches to AI's civil liability,the structural approach applies the traditional liability types to AI,including contractual liability,product liability,and tort law.In terms of contractual obligations,the European states mainly stipulated that if the seller failed to perform his obligations,the buyer had the right to claim damages,but the seller was only responsible for the result of a possible non-performance that he foreseen or could reasonably expect.In the United States,manufacturers are responsible for the injury of personsand property caused by breach of warranty,which warranty may be either express or implied;In the area of product liability,the manufacturer is responsible for damage to the buyer due to defects in his products.Product defects may be: a)manufacturing defects;b)design defects;c)warning defects.However,due to the high degree of autonomy of artificial intelligence,it is hard to believe that it is possible to demarcate the boundary causing damages:the will of the AI,that is,from its own decision or from the defect of the product itself.Therefore,it is very hard for the plaintiff to prove that AI has product defects.In the area of tort law,a better proposal now is to apply the principle of strict liability,using AI as an "abnormally dangerous activity" in some settings.In a functional approach,the risk and precaution of AI and the direct relationship among elements of civil liability is discussed.AI-related risks are included in a particular category of risks: the risk of unease.These types of risks are unreasonable,though not yet scientifically proven.Therefore,this risk is suspicious but not yet established.In order to prevent the risk of AI,it is necessary to evaluate the risk to guide the supervision of AI.The EU adopts the precautionary principle.However,applying the precautionary principle to AI can also lead to over-protectionism that can hinder AI's innovation.However,the precautionary principle may be useful in assessing the causation between AI damage and under the precautionary principle it is important to include a clause that reverses the burden of proof and entitles the producer,manufacturer or importer to take the burden of proof.In the case of many manufacturers involved in the production of AI,the manufacturers should bear joint and several liability.Internally,based on the prevention logic,damage may be proportionally shared among manufacturers based on the precautions they should take.Based on the above analysis,the suggestion to deal with the civil liability of artificial intelligence in our country is based on the existing "Contract Law","Product Quality Law","Consumer Protection Law" and "Tort Law" to find victims the justified basis for the right to request.When artificial intelligence develops completely autonomy,it should be separately legislated,given its independent legal status,and establish a system of responsible property.