Beijing has taken its most direct regulatory swing yet at AI-enabled harassment, with the Cyberspace Administration of China (CAC) publishing a draft cyberbullying law on Wednesday that explicitly targets content generated or amplified by artificial intelligence. The proposed rules go further than previous CAC guidance by requiring platforms to actively trace, flag, and remove AI-generated abuse, placing the enforcement burden squarely on the technology companies that host China’s 1.1 billion internet users.
The draft law is open for public comment until August 28, giving the tech industry and civil society a narrow window to respond before the rules are finalized. It follows a series of increasingly assertive regulatory moves by the CAC, which has been systematically building a framework for governing AI-generated content since 2022.
Key Provisions of the Draft Law
The draft law outlines several key requirements for online platforms and users. First, it establishes extraterritorial jurisdiction: the law would apply to cyberbullying activities carried out within China, as well as to organisations and individuals overseas that target the country. Second, it mandates real-name verification: online platforms would be required to verify the real identities of users before allowing them to post information or use instant messaging services. Third, and most technically demanding, the proposed rules ask platforms to trace and report the risk of cyberbullying content generated or spread using AI technology, remove and block such content, and report it to authorities. Schools are also required to incorporate anti-cyberbullying material into their curricula.
On the enforcement side, online service providers that violate the rules could face fines of up to 10 million yuan ($1.5 million), the shutdown of their websites or applications, or the revocation of business licenses.
The Challenge of AI-Enabled Abuse
The specific inclusion of AI-enabled abuse highlights the Chinese government’s recognition of the unique threats posed by generative technologies. As AI tools become more accessible and sophisticated, the ability to generate realistic deepfakes, synthetic voice clones, and automated harassment campaigns has increased dramatically. The same AI capabilities that have powered China’s booming digital human and AI video industries, sectors EastFrontier has covered extensively, including Beijing’s earlier moves to regulate AI chatbots with new draft rules, are now being turned toward malicious ends.
By requiring platforms to actively trace and report AI-generated cyberbullying content, the CAC is placing the burden of enforcement squarely on the tech companies that host these communities. This aligns with China’s broader approach to internet regulation, which heavily relies on platform-level moderation and compliance. The real-name verification requirement reinforces the state’s ability to identify and penalize individuals engaging in online abuse.
However, the technical challenge of accurately identifying and tracing AI-generated content remains significant. Platforms will likely need to invest heavily in new detection tools to comply with the proposed rules — a requirement that could disproportionately burden smaller operators while entrenching the advantages of the large incumbents who already have sophisticated content moderation infrastructure. As the public consultation period progresses, the tech industry will be watching closely to see how these regulations are finalized and what compliance timelines are ultimately imposed.
The draft law also raises broader questions about the global governance of AI-enabled harm. China is not alone in grappling with this challenge: the European Union’s AI Act, the UK’s Online Safety Act, and various US state-level proposals all attempt to address AI-generated abuse in different ways. What distinguishes the CAC’s approach is its emphasis on platform-level traceability, the requirement to maintain logs of AI-generated content that can be surfaced on demand by regulators.
This creates a surveillance infrastructure that serves dual purposes: protecting users from abuse while also giving the state visibility into the AI content ecosystem. Whether that trade-off is acceptable will be a matter of ongoing debate, both within China and in the international community watching how Beijing’s AI governance model evolves.
For international observers, the draft law is also a data point in a larger question: whether China’s AI regulatory framework is converging toward global norms or diverging from them. The CAC’s emphasis on platform accountability, real-name verification, and state-accessible content logs reflects a governance philosophy that prioritizes social stability and state oversight over individual privacy. As Chinese AI platforms expand internationally, the question of whether they will be required to implement the same traceability infrastructure in foreign markets, and whether foreign regulators will accept or reject that architecture, will become increasingly consequential.
The draft law’s public comment period, which runs until August 28, will be an important signal of how much room the CAC is willing to give the tech industry to shape the final rules. In previous regulatory cycles, Chinese internet companies have successfully lobbied for modifications to technically burdensome requirements, but the political priority attached to cyberbullying, particularly following high-profile cases involving minors, suggests that the core provisions are unlikely to be substantially weakened.