Abstract:"Deep fake" mainly refers to the surreal digital forgery of images, video and audio.As a branch of artificial intelligence, the unique bionic skills of "deep fake" are increasingly emerging in fields such as health care and education, but they also pose risks to personal information, women's reputations and property rights, and the public's right to know and the system of preexisting evidence.n addition to technological innovation, we should be cautious about the development and application of "deep forgery" technology.Weigh the technical risks, stakeholders such as Internet operators and managers involved, and public interests.On this basis, the following measures should be taken: First, from the legal perspective, it is not appropriate to expand the existing circle of illegal crimes, and abide by the nature of legal modesty.Secondly, technically speaking, "self-certification" is combined with "other certification".On the one hand, it requires "deep forgery" of the producer to identify the attribute of the work;On the other hand, a collaborative effort to develop a detection timestamp tool.Thirdly, the idea of normalized stage management needs to be implanted at the platform level.The above approaches are the proper meaning of the concept of collaborative governance, and also the way to maintain the balance between scientific and technological innovation and the interests of data subjects.
