中國深偽技術（Deepfake）法令規範之研究

Research on the Regulations and Norms of Deepfake Technology in China

隨著AI人工智慧的發展，世界各國面對深偽技術（Deepfake）的科技已經感受到威脅，「早就針對深偽影片修訂法律，並發展AI測偽技術，我國之前對深偽技術的影片很少正視，直到網紅朱玉宸（小玉）的名人變臉謎片遭破獲，官方才醒了過來。」我國直到2023年2月8日才公布增訂《刑法》第10條第8項明確規範性影像之定義及《刑法》第28章之1妨害性隱私及不實性影像罪。
反觀中國對於深偽技術（Deepfake）法令規範在《民法典》、《刑法》與行政法令上已經日趨完善，包含《民法典》第1019條侵害他人肖像權；《刑法》第363條淫穢物品牟利犯罪、《刑法》第363條傳播淫穢物品罪、《刑法》第246條誹謗罪、《刑法》第433條利用造謠惑眾罪、造謠擾亂軍心罪；《著作權法》第10條表演權、第39條表演者權利、第40條表演形象不受歪曲的權利；《個人信息保護法》第51條個人信息處理保護措施、第57條個人信息處理者的補救與通知措施、第66條違反個人信息之罰款；《關於加強網絡信息保護的決定》第8項至第11項有權要求網絡服務提供者刪除信息權、舉報權、網絡服務提供者應配合提供信息、網絡服務業務之處罰等；《網路音視頻信息服務管理規定》第11條信息服務使用者不得傳播假信息；《網路信息內容生態治理規定》第23條網絡信息內容服務使用者和網絡信息內容生產者、網絡信息內容服務平台不得利用深度學習從事違法活動及《具有輿論屬性或社會動員能力的互聯網信息服務安全評估規定》第3條第2款互聯網信息服務提供者應當開展安全評估並對評估結果負責等規定。
本文從深偽技術定義、產生方式與中國深偽技術法令規範之研究結果，拋磚引玉提出中國修正《刑法》及我國《民法》增修條文草案、《著作權法》增修條文草案及增訂《網路影音資訊服務管理法》草案，提供未來深偽技術（Deepfake）規範上之修法建議。

With the development of AI artificial intelligence, countries around the world have felt a threat to the technology of Deepfake technology, We have long revised laws and developed AI anti-counterfeiting technology for Deepfake films. Previously, our country rarely faced up to films with Deepfake technology, and it wasn’t until the celebrity face changing mystery film of internet celebrity CHU, YU-CHEN (HSIAO-YU) was cracked that the authorities woke up. It wasn’t until February 8, 2023 that ROC announced and revised the definition of normative images in Article 10(8) of the Criminal Code, as well as the crime of disturbing sexual privacy and false images in Chapter 28 of the Criminal Code.
On the other hand, mainland PRC has gradually improved its regulations on Deepfake technology in Civil Code, Criminal Code, and Administrative Decree, including Article 1019 of the Civil Code, which infringes on the right to portrait of others; Article 363 of the Criminal Code commits the crime of profiting from obscene materials, Article 363 of the Criminal Code commits the crime of spreading obscene materials, Article 246 of the Criminal Code commits the crime of defamation, Article 433 of the Criminal Code commits the crime of using rumors to mislead the public, and Article 363 of the Criminal Code commits the crime of spreading rumors to disturb the morale of the military; Article 10 of the Copyright Law is the right to perform, Article 39 is the right to perform, and Article 40 is the right not to be distorted in the image of performance; Article 51 of the Personal Information Protection Law: Personal Information Processing Protection Measures, Article 57: Remedies and Notification Measures for Personal Information Processors, and Article 66: Penalties for Violation of Personal Information; Items 8 to 11 of the Decision on Strengthening the Protection of Network Information have the right to require network service providers to delete information, report information, cooperate with network service providers in providing information, and impose penalties on network service businesses; Article 11 of the Regulations on the Management of Online Audio and Video Information Services prohibits information service users from spreading false information; Article 23 of the Regulations on the Ecological Governance of Network Information Content stipulates that users of network information content services, network information content producers, and network information content service platforms shall not engage in illegal activities through deep learning, and Article 3, paragraph 2 of the Regulations on the Security Assessment of Internet Information Services with Public Opinion Properties or Social Mobilization Ability stipulates that internet information service providers shall conduct security assessments and be responsible for the assessment results.
Based on the definition and generation of Deepfake technology and the research results of PRC’s Deepfake technology laws and regulations, this paper puts forward the draft of amendments to the PRC’s Criminal Code and the draft of amendments to the ROC’s Civil Code and the Copyright Law and the draft of amendments to the Internet Audio Video Information Service Management Law, providing suggestions for future Deepfake technology regulations.