China enacted its latest regulations on artificial intelligence (AI)-generated content this week. The new rules, which are a watered-down version of stricter draft regulations, aim to keep China competitive in the AI field while maintaining strict censorship on online content.
The rapid advancements in generative AI have raised concerns worldwide about the technology’s potential for disinformation and misuse, especially with the rise of deepfake images that manipulate and alter visual content.
Chinese companies have been eager to develop AI services that can mimic human speech, especially since the release of San Francisco-based OpenAI’s ChatGPT, which is banned in China.
The 24 new regulations are seen as less stringent than the earlier draft regulations, suggesting that Beijing is looking to encourage domestic players in the AI industry, which is currently dominated by the US.
Key Points About Beijing’s Regulations
1. AI Ethics
Generative AI must follow the core values of socialism and refrain from threatening national security or promoting terrorism, violence, or ethnic hatred. Service providers must label AI-generated content accordingly and take measures to prevent gender, age, and racial discrimination in algorithm design. The software should not create content that contains false or harmful information. Additionally, AI programs must be trained using legally obtained data sources that respect intellectual property rights, and individuals must provide consent before their personal information is used in AI training.
2. Safety Measures
Companies offering publicly available generative AI software must take effective measures to prevent excessive reliance or addiction among underage users. They are also required to establish reporting mechanisms for the public to report inappropriate content and promptly delete any illegal content. Service providers must conduct security assessments and submit filings on their algorithms to authorities if their software is deemed to impact public opinion. This is a step back from earlier draft rules that required security assessments for all public-facing programs.
3. Enforcement
The regulations are considered provisional measures and are subject to existing Chinese laws. They are the latest in a series of regulations that target various aspects of AI technology. While an earlier draft suggested fines of up to 100,000 yuan ($13,824) for violations, the latest version states that violators will receive warnings or face suspension. More severe punishment will be enforced only if they are found to be in breach of actual laws.
4. Supporting Innovation
The new regulations indicate a shift towards supporting development and innovation in the AI industry. While the earlier draft aimed to maintain strict control over online content, several restrictions on generative AI have been softened. The scope of the rules now applies only to generative AI programs available to the public, excluding research and development uses. This suggests that Beijing recognizes the need to compete in the global AI race.
China’s approach to AI legislation falls between the strict approach of the EU and the more lenient approach of the United States. By encouraging the development of domestic AI technology, China aims to maintain its competitiveness in the rapidly evolving AI industry.