China introduces rules for generative AI services like ChatGPT

Chinese regulators on Thursday finalized first-of-a-kind rules for generative artificial intelligence, as the country looks to increase oversight of the fast-growing technology.

The powerful Cyberspace Administration of China (CAC) said it was working with several other regulators to come up with the new regulation, which will take effect on August 15.

Generative AI is a rapidly growing area of ​​technology where artificial intelligence services are able to generate content such as text or images. ChatGPT, developed by the American firm OpenAI, is the most famous example and allows users to ask the chatbot and receive answers to questions.

These services are trained on huge amounts of data. However, ChatGPT’s success has sparked a wave of rival services emerging, raising concerns among global regulators about the potential risks surrounding the technology.

Chinese tech giants have jumped on the bandwagon, announcing plans and launching their own generative AI services.

But China, which tightly controls the domestic internet via censorship and regulation, is watching AI developments closely. Chinese regulators are concerned about the potential for these services to generate content that may conflict with Beijing’s views or ideology.

That’s partly why Chinese tech firms have been cautious about launching their ChatGPT-like services. Instead of complete services that are widely available to the public, Chinese companies have focused their technology on enterprise and narrow uses.

For example, this month Alibaba launched an artificial intelligence tool that can generate images from messages called Tongyi Wanxiang, but it is only available to corporate customers for beta testing.

However, the rules from the CAC give the technology giants a framework to work with when it comes to technology.

The rules will only apply to generative AI services that are available to the general public rather than those that are developed in, for example, research institutions.

Generative AI services must obtain a license to operate, the CAC said.

If a generative AI service provider finds “illegal” content, it should take steps to stop generating that content, improve its algorithm, and then report that material to the relevant authority.

Providers of these services must conduct security assessments of their product and ensure that user information is secure.

Generative AI services in China must also adhere to the “core values ​​of socialism,” the CAC said.

Still, regulators are trying to find the balance between making China a leader in artificial intelligence while keeping an eye on developments.

CAC’s rules said the regulation aims to encourage innovative applications of generative AI and support the development of related infrastructure such as semiconductors.