The Cyberspace Administration of China (CAC) Tuesday released a draft law detailing new regulations for providers of artificial intelligence (AI) products. The new law will require these providers to report to the Ministry of Cyberspace affairs and ensure their product undergoes a security assessment before it can be released to the public.

Although all AI products developed in China will need to be assessed before public use, the law also outlined that the state supports the “independent innovation, popularisation and application” of AI products. The CAC “encourages use of safe and reliable [AI] software”, however, their new regulations state that such software must generate content that “reflects the core values of socialism.” It must also “not contain any content that subverts state power, [or] advocates for the overthrow of the socialist system.”

Among these new requirements, the CAC regulations set out articles protecting user information. Providers are now obliged to protect the user’s input and are not allowed to illegally retain any input information. The AI products are also banned from generating any discriminatory content based on the user’s race, country, gender and other characteristics. Furthermore, providers are required to create a suitable method of receiving and dealing with user complaints about their product.

These legal provisions follow a recent increase in use of AI chatbots and image generators such as ChatGPT, a widely used chatbot created by Open-AI in November 2022. Most recently, in March, Italy’s Data Protection Authority (DPA) banned ChatGPT, citing the lack of an age verification system for children users and unlawful collection of users’ personal data. Large Chinese technology companies such as Alibaba and SenseTime, revealed their newest AI models in recent weeks, which will likely be subject to the CAC’s new regulations.