China initiates a campaign lasting several months to combat the misuse of AI.

      The annual 'Qinglang' campaign by the Cyberspace Administration is entering a markedly different regulatory landscape compared to last year's initiative, coinciding with the White House's allegations of China conducting large-scale AI theft operations. According to Reuters, China has initiated a months-long enforcement campaign focused on the misuse of artificial intelligence.

      This campaign, launched by the Cyberspace Administration of China (CAC) in coordination with the Ministry of Public Security and other agencies, is geared towards combating AI-enabled fraud, deepfakes, disinformation, and illegal applications that infringe on privacy and intellectual property rights. This action marks the 2026 iteration of the annual 'Qinglang' (Clear and Bright) campaign series. Its immediate predecessor, named 'Rectification of AI Technology Misuse,' was launched on April 30, 2025, and lasted three months, divided into two phases.

      By the conclusion of its first phase in June 2025, authorities had removed over 3,500 AI-related products, deleted more than 960,000 pieces of illegal or harmful content, and closed or sanctioned more than 3,700 accounts. This year’s campaign is set against a much more advanced regulatory landscape and a geopolitically charged atmosphere, making its objectives and targets significantly more complex than those of its predecessor.

      What does the campaign focus on?

      China’s enforcement actions against AI misuse are organized around categories of misuse that have broadened with each iteration, in line with advancements in AI capabilities and their criminal applications. Following the established Qinglang enforcement framework and new regulations introduced in 2025 and early 2026, the current campaign is anticipated to target numerous categories simultaneously.

      The first and most commercially significant category is AI-enabled fraud and impersonation. There has been a notable rise in the use of deepfake technologies for voice-cloning and face-swapping to impersonate celebrities, business leaders, and government officials in scams directed at ordinary consumers. The CAC's 2025 campaign specifically addressed the employment of AI to impersonate relatives and friends for online fraud and the improper use of AI to create likenesses of deceased individuals without consent.

      On April 3, 2026, the CAC issued draft regulations for digital virtual human services, which include consent requirements for likeness usage and prohibitions on bypassing biometric authentication systems. The second major focus is AI-generated disinformation and ‘online water army’ activities, characterized by the industrial-scale use of AI to develop fake social media accounts, produce and distribute coordinated content, influence engagement metrics, and establish artificial trending topics. The 2025 campaign identified this as a critical area for its second phase, concentrating on platforms that enable AI-powered account farming and social bot networks.

      The third category pertains to non-compliance with mandatory filing and registration processes. China mandates that large language models (LLMs) offering generative AI services to the public undergo a security assessment and secure filing with the CAC before their launch. By March 2025, only 346 generative AI services had completed this filing, with many others still unregistered. The first phase of the 2025 campaign highlighted unfiled AI products as a key target, leading local regulators to penalize three AI applications for providing services without completing the necessary processes.

      The fourth focuses on the management of training data, specifically utilizing training datasets that contain content violating intellectual property rights, privacy rights, or consent requirements. This aspect of enforcement is particularly sensitive in 2026, especially following the White House's formal accusation on April 23 linking Chinese companies to large-scale campaigns aimed at extracting capabilities from U.S. frontier AI models using jailbreaking techniques and thousands of proxy accounts.

      China's domestic enforcement initiative does not directly address this U.S. accusation; rather, it concentrates on safeguarding the rights of Chinese holders and users. However, both regulatory ecosystems are evolving with an awareness of one another. The 2026 campaign benefits from a significantly more established domestic regulatory framework than its predecessor. Several important regulations were either enacted or published as drafts in the months preceding this enforcement action.

      On September 1, 2025, China’s mandatory AIGC (AI-generated content) labeling standards came into effect, requiring visible and technical labels for all AI-generated text, images, audio, and video. On April 10, 2026, the CAC released the Interim Measures for the Management of Anthropomorphic AI Interactive Services, governing chatbots and AI companions, which will take effect on July 15, 2026. Additionally, the CAC published draft rules concerning biometric deepfakes on April 3, 2026, with a public comment period ending on May 6, 2026. Also in April 2026, a collaborative enforcement agenda targeting personal information protection across seven sectors, including internet advertising and healthcare, was released by the CAC, MIIT, and MPS.

      The cumulative effect of this regulatory development is that the 2026 Qinglang campaign possesses considerably more legal authority compared