China Drafts Rules To Manage AI Services, Firms May Need To Provide Security Assessments

Providers will be accountable for the legitimacy of data used to train generative AI products and must take steps to prevent algorithmic discrimination.

If unsuitable content is produced, platforms must be updated within three months.

China’s Cyberspace Administration has released draft regulations on managing generative artificial intelligence (AI) services, requiring that companies provide security assessments to authorities before launching their products. With governments around the world seeking ways to mitigate the potential risks of emerging technologies, the move comes after a recent surge in investment and consumer popularity of AI products in China.

It also follows demonstrations by major Chinese tech firms, including Alibaba, SenseTime, and Baidu, of new AI models that can support chatbots and image generators. While the CAC acknowledged China’s support for AI innovation and encouraged the use of reliable software, tools, and data resources, it warned that AI-generated content must adhere to the country’s core socialist values.

Providers will be accountable for the legitimacy of data used to train generative AI products and must take steps to prevent algorithmic discrimination. The CAC also requires service providers to mandate the submission of real user identities and related information. Failure to comply with regulations could lead to fines, suspended services, and even criminal investigations.

If unsuitable content is produced, platforms must be updated within three months to avoid similar content being generated again. Public comment on the proposals is invited until May 10th, with the regulations expected to be implemented later this year.

Meanwhile, OpenAI's viral AI chatbot ChatGPT has been blocked in Italy, the media has reported. Italy's data protection agency has opened ChatGPT probe over privacy concerns due to a suspected breach of data collection rules. 

The country's data protection watchdog has also accused ChatGPT of failing to check the age of its users, which should be reserved for people aged 13 and above. The agency said it had provisionally restricted chatbot's use of Italian users' personal data.

The Italian Data Protection Authority said that the decision "with immediate effect" will result in "the temporary limitation of the processing of Italian user data vis-a-vis OpenAI." 

Italy's privacy watchdog has said it will immediately block and investigate OpenAI, the US-based company behind the viral AI tool, from processing the data of Italian users. The order is temporary until the company respects the EU's landmark privacy law, the General Data Protection Regulation (GDPR).