China- August 2023: China has enacted its latest regulations on AI-generated content, signaling a strategic shift towards fostering innovation while maintaining a firm grip on online censorship. The new rules, implemented this week, are a moderated version of earlier drafts that sought stricter control. The revisions reflect Beijing's dual priorities: encouraging domestic advancements in artificial intelligence to compete globally while upholding stringent oversight of digital content.
Generative AI technology, capable of producing human-like speech and deepfake imagery, has raised international concerns about its misuse for disinformation. China’s response to this global challenge includes a set of 24 provisions targeting public-facing AI services. These measures emphasize ethical compliance, content transparency, and user safety.
AI services in China must align with socialist values and avoid promoting content that could jeopardize national security or incite violence and discrimination. The guidelines mandate clear labeling of AI-generated content, ethical algorithm design, and reliance on legally obtained training data. Consent is also required for using personal information in AI training, ensuring the protection of intellectual property and privacy rights.
Safety measures include mechanisms to address misuse and prevent addiction among minors. Companies are required to allow public reporting of inappropriate content and must delete unlawful material promptly. Developers of generative AI software are obligated to conduct security assessments and file reports with regulators if their programs influence public opinion. However, the requirement for comprehensive security checks across all public-facing AI systems, a feature of the earlier draft, has been scaled back to facilitate innovation.
The rules, classified as provisional measures, coexist with broader Chinese laws governing digital and online activities. Unlike the European Union’s stringent approach or the United States’ comparatively lenient stance, China’s regulatory framework reflects a targeted strategy, addressing specific facets of AI technology. This vertical focus allows for nuanced oversight tailored to rapidly evolving technologies like generative AI.
Experts suggest that the softened measures underline China’s intent to secure its position in the global AI race. Earlier drafts proposing fines up to 100,000 yuan ($13,824) for violations have been replaced with milder penalties such as warnings or suspensions unless broader laws are breached.
This regulatory evolution signals a recalibration of priorities. While maintaining robust censorship remains central, the rules have shifted to encourage technological progress. Restrictions on research and development applications, for instance, have been relaxed, with the focus narrowed to AI systems available to the general public. Observers interpret this as a recognition of AI’s strategic importance, with Beijing striving to balance innovation with control.
Jeremy Daum, a Senior Fellow at Yale Law School’s Paul Tsai China Centre, noted that the revised rules aim to support AI industry growth without compromising governmental oversight. "China appears to embrace the notion of an AI race, adjusting its stance to remain competitive globally," he observed. This pragmatic approach reflects China’s ambition to lead in AI development while navigating the complex interplay of innovation and regulation.