As a "container of human intelligence", AI tool has become a popular traffic password with its powerful functions of deep learning, data analysis, language processing, image generation, video simulation, etc. Recently, a woman's subway photo was released and widely circulated online using AI software "one click undress", reflecting the urgent need to address the threat and chaos caused by malicious research and abuse of AI tools. The "AI one click undressing" function can be foreseen to result in infringement of personality rights, copyrights, and other aspects during the development stage. The use of deep synthesis technology to preserve the head, replace the body, and fabricate portraits can infringe upon the right to the portrait; The act of generating false and indecent images and triggering online dissemination will infringe on reputation rights; The generated image contains the shooting elements of the original photo, and even constitutes a substantial similarity to the original photo, infringing on the copyright of the original photo author. This function seriously violates social order, good customs, and socialist core values, and should be judged as not to be available in the early stages of development. However, under strict control, there are still cases of crossing the line. Although malicious software and technologies such as AI facechanging and AI stripping have long been banned in China, there are still foreign developed software or alternative software with similar functions that have entered the domestic market. The author believes that to eliminate the chaos of malicious development and abuse of artificial intelligence and promote the healthy development of artificial intelligence technology, it is necessary to achieve multi-party efforts and diversified governance. On the one hand, we need to strengthen supervision. Strictly implement requirements such as registration, security assessment, and prominent labeling for deep synthesis service providers and technical supporters; The Ministry of Industry and Information Technology should strengthen the review and spot check of AI apps, mini programs, fast applications, and other programs or software on the shelves, and take timely technical measures to block and prohibit the downloading and use of illegal software imported from abroad; The Cyberspace Office aims to cut off the transmission channels of AI infringing works and create a clean and upright online environment through the "Qinglang" series of special actions. On the other hand, we need to reconstruct the tort liability system. Clarify the infringement subjects, liability structures, and attribution principles under various forms of generative AI, such as foreseeable infringement during the research and development stage, automatic infringement during AI tool operation process, and infringement after receiving instructions from natural persons, in order to implement judicial accountability. In addition, we should use hot cases and news information such as "AI Companion Infringement Case," "AI Face Changing Case," and "AI Video Fraud Case," which have attracted widespread attention from society, to carry out case studies and guide netizens to overcome blind obedience and curiosity, carefully identify and use AI tools, and avoid being secretly collected by unscrupulous developers, excessively requesting data permissions, and thus infringing on their legitimate rights. (New News Agency)