AI Legislation
China proposed rules governing artificial intelligence that span data, computing, algorithms, deepfakes, and privacy.
14 on file · 12 became law · 2 still moving · 1 moved in 30 days
China proposed rules governing artificial intelligence that span data, computing, algorithms, deepfakes, and privacy.
Draft rules that would designate handlers processing personal information of 10 million or more people or providing important network services as large handlers, subject to registration, heightened security duties and independent personal information protection supervisory committees.
Regulates AI companion and other human-like emotional interaction services, barring virtual intimate relationship services for minors and requiring safeguards against emotional dependence and separate consent before chat data is used for training.
Amends the Cybersecurity Law to add a provision supporting AI research, infrastructure, ethics and risk monitoring, raise fines for network security violations, and align liability with the Data Security Law and Personal Information Protection Law.
Sets out the certification route for transferring personal information out of China, including which handlers may use it, the certification bodies and process, and supervision of certified transfers.
Requires providers of AI generation services and content distribution platforms to add explicit visible labels and implicit metadata labels to AI-generated text, audio, images, video and virtual scenes, and bars removing or forging those labels.
Sets the mandatory national technical standard for how explicit and implicit labels, including metadata, must be applied to AI-generated text, images, audio, video and virtual scenes under China's labeling measures.
Sets China's national plan to integrate AI across science, industry, consumption, public services, governance and international cooperation, targeting over 70 percent adoption of intelligent terminals and agents by 2027 and over 90 percent by 2030.
Creates a voluntary state-run online identity system that issues citizens a network ID number and network certificate so they can verify their identity to platforms without handing over ID card details, while platforms must keep other verification options.
Requires personal information handlers to conduct regular compliance audits, with those processing data on more than 10 million people auditing at least every two years, and lets regulators order audits when risks arise.
Implements the Cybersecurity Law, Data Security Law and Personal Information Protection Law with detailed rules on personal information handling, important data protection, cross-border data transfers and obligations of large online platforms.
Relaxes China's data export rules by exempting several categories of cross-border transfers from security assessment, standard contract or certification requirements and letting free trade zones set negative lists.
Sets rules for providers offering generative AI services to the public in China, covering training data legality, content controls, labeling, user protection, security assessments and algorithm filing.
Regulates deep synthesis technologies such as face swapping and voice cloning, requiring providers to verify users, label synthetic content, obtain consent for editing biometric features and conduct security assessments.