On Tuesday, China’s powerful Cyberspace Administration (CAC) released draft rules to manage the development of generative artificial intelligence (AI) products like ChatGPT, which have become increasingly popular in the country. The rules are the first of their kind and are designed to target the fast-growing field of generative AI, as domestic tech giants begin rolling out their own ChatGPT-style products.
What is Generative AI?
Generative AI refers to algorithms trained with vast amounts of data that can generate content like images and text. OpenAI developed ChatGPT, which is capable of generating responses to user queries and has become hugely popular.
Chinese Tech Giants Unveil Rival Generative AI Products
In recent weeks, Chinese tech giants have been introducing their own competing generative AI products. On Tuesday, Alibaba unveiled its own generative AI product, Tongyi Qianwen, which the e-commerce giant plans to integrate across various services. Baidu launched its equivalent, Ernie Bot, for testing last month.
CAC’s Draft Measures for Generative AI
The draft measures from the CAC specify the guidelines that generative AI services must adhere to, which includes defining the type of content these products can produce. According to the draft rules, the content must reflect the core values of socialism and should not subvert state power. According to the regulator, companies are required to ensure that the data used for training AI models does not exhibit any discriminatory behavior based on factors such as ethnicity, race, and gender. Furthermore, generating false information is also prohibited, as stated by the regulator.
Political Sensitivity
As per the views of analysts shared with CNBC earlier, Chinese regulators are expected to be closely monitoring the development of generative AI due to its potential to generate politically sensitive content. The CAC’s rules address this concern and provide a framework for Chinese companies to follow when developing such technology.
Other Regulations on Data Protection and Algorithm Development
The measures outlined by the CAC are scheduled to be implemented later this year, and they will complement other regulations in China related to data protection and algorithm development. Although China is not the sole nation worried about the emergence of generative AI, Italy also banned ChatGPT in March due to privacy concerns.
Conclusion
The CAC’s draft rules are designed to regulate the development of generative AI products like ChatGPT in China. These measures require that the content generated reflects the core values of socialism and should not undermine state power. Additionally, data used to train these AI models must not discriminate against people based on race, gender or ethnicity and should not generate false information. The regulations are expected to come into effect later this year and will work in conjunction with other Chinese regulations on data protection and algorithm development.
