- Chinese regulators on Tuesday released draft rules designed to govern how companies develop so-called generative AI products such as ChatGPT.
- China’s draft Cyberspace Management Measures sets out the basic rules that generative AI services must follow, including what kind of content these products are allowed to create.
- For example, content generated by AI must reflect the core values of socialism and must not undermine state power, according to the draft rules.
Chinese tech giants from Baidu to Alibaba have launched their own ChatGPT competitors. These products are examples of what is called generative artificial intelligence (AI). These are artificial intelligence services that can generate images or text after user queries. But artificial intelligence also worries regulators. China’s powerful Cyberspace Administration has released draft rules governing how generative AI products should be developed.
CFOTO | Publishing in the future | Getty Images
Chinese regulators on Tuesday released draft rules designed to govern how companies develop generative AI products such as ChatGPT.
The draft rules from China’s powerful cyberspace administration are the first of their kind in the country and target rapidly evolving artificial intelligence as local tech giants begin rolling out ChatGPT-style products.
So-called generative AI refers to algorithms trained on huge amounts of data that are capable of creating content such as images and text. The American company OpenAI developed ChatGPT that is able to generate responses to user inquiries and it has become very popular.
Over the past few weeks, Chinese tech giants have unveiled their competing products. On Tuesday, Alibaba unveiled Tongyi Qianwen, its generative AI product, which the e-commerce giant plans to integrate across various services. Last month, Baidu also launched its equivalent, the Ernie Bot, for testing.
The CAC’s draft measures lay out the ground rules that generative AI services must follow, including what type of content those products are allowed to create. The content should reflect the core values of socialism and should not subvert state power, according to the draft rules.
The CAC said companies should ensure that the data used to train these AI models will not discriminate against people based on things like race, ethnicity, and gender. The authority added that it should not generate false information.
Analysts previously told CNBC that Chinese regulators will likely be keeping a close eye on the development of generative AI given the potential to produce potentially politically sensitive content.
The CAC rules highlight this concern and set a framework for how Chinese companies need to approach technology development. But the measures, which are set to take effect later this year, will work in conjunction with various other Chinese regulations around data protection and algorithm development.
However, China is not the only country concerned with developing generative AI. In March, Italy banned ChatGPT due to privacy concerns.
“Unapologetic reader. Social media maven. Beer lover. Food fanatic. Zombie advocate. Bacon aficionado. Web practitioner.”