Research On Tort Law In AI Face Changing Application

In recent years,AI face-changing technology has been constantly evolving and developing,which provides convenience for life and saves the production cost of film and television works.However,in real life,this technology is used in a large number of pornographic websites and pornographic videos without permission,which seriously infringes on the right of portrait,copyright and other rights of the party concerned.Although this technology has just started to develop in China,facing the potential risks that will explode soon after the emergence of this technology,the legal community must have a forward-looking prediction and response mechanism,and it is urgent to strengthen the supervision of AI face changing technology.When the infringement occurs,only by correctly identifying the infringement and applying the law to solve the infringement problem of AI face-changing technology,can the benign development of science and technology be guaranteed.AI face changing technology may cause many infringement problems.In terms of right determination,when the direct actor replaces the portrait of others into the video of film and television works without the permission of others,such behavior constitutes a violation of portrait right.However,in some cases,the direct actor may use the whole face or only half of the face to carry out the tort,as well as the "deep learning" technology to use the portrait of "similar looking people" to carry out the tort.Whether these acts constitute infringement needs to be determined by specific identifications.In addition,the current law on portrait right infringement does not make clear whether the network communication behavior infringes portrait right or not,which makes it difficult to identify the behavior of using information network to spread others' portrait in practice.It is more beneficial to protect the rights of victims to include the network communication behavior.In the film and television works,there are still some problems such as whether the image right of the replaced person is infringed when half of the face or part of the facial features are replaced,whether the face changing behavior constitutes the identification of the distortion of the performance image,whether the distortion of the virtual character image constitutes the identification of infringement,and the public figure's tolerance obligation in the AI face changing.This paper proposes that: first of all,whether the use of half a face or part of a facial feature infringes the right of portrait of the performer should be recognized as the standard of the replaced person.Secondly,the distortion of performance image should be based on the premise of subjective malice.Thirdly,the distortion of virtual character image may constitute a violation of the right of modification and the right to protect the integrity of the work.Finally,public figures should have a tolerance obligation in certain cases of AI face-swapping.In the aspect of liability,as long as the direct actor,as a tort,has committed an act that infringes upon the rights of others,he shall be held responsible.Network service provider is generally applicable fault liability imputation principle,however,network service providers in our country is not like America in face identification technology,enable the AI because AI technology in face of professional and technical,the patentee,it is often difficult to prove the existence of network service provider fault for network service provider fails to perform the obligations of the situation of the direct infringement applicable no-fault principle is more reasonable to solve.In addition,software developers are usually not liable based on the "technology neutrality principle".However,when software developers intentionally force network users to give up their right of portrait through user agreement,they should share the infringement with the direct actor or Internet service provider.