Tencent: China’s New Rules for ChatGPT-like AI Chatbots Include ‘Socialist Values’
China has recently released a comprehensive set of regulations aimed at governing generative artificial intelligence (AI) models, with a strong emphasis on upholding core socialist values. These regulations will only apply to AI services offered to the public, as Chinese regulators seek to promote the development of this evolving technology.
The regulations, developed by a consortium of seven Chinese regulatory bodies led by the Cyberspace Administration of China, are set to take effect on August 15. The regulators have stated their intention to adopt an inclusive and cautious approach towards generative AI services, implementing a graded regulatory approach to ensure effective oversight.
Under these regulations, all generative AI content services available to the Chinese public, including text, pictures, audio, and video, will be subject to these rules.
Companies providing generative AI services are now required to adhere to core socialist values. These services should refrain from generating content that instigates subversion of state power or the overthrow of the socialist system. Additionally, they must avoid endangering national security and interests, damaging the country’s image, inciting secession or undermining national unity and social stability. The regulations also prohibit generative AI models from promoting terrorism, extremism, national hatred, ethnic discrimination, violence, obscenity, and pornography.
Importantly, AI models and chatbots are expected to avoid generating misinformation that could be harmful.
The regulations stipulate that algorithm design and training data selection must not be based on discrimination regarding ethnicity, faith, country, region, gender, age, occupation, or health. Furthermore, respect for intellectual property rights is required. The new rules also aim to encourage innovation in the application of generative AI technology and the creation of positive and healthy content.
The final regulations demonstrate a more positive tone than the draft version released in April. Punitive measures mentioned in the earlier draft, including fines ranging from 10,000 to 100,000 yuan for technology-related offenses, have been omitted.
However, the regulations make it obligatory for generative AI service providers to cease operations and report any unlawful activities if they become aware that users are employing their AI services to generate illegal content.
In addition, AI service providers must implement safeguards to prevent minors from developing addictive habits towards their services.
Chinese regulators have assured their support for and encouragement of the development of this new technology through appropriate measures.
As of now, Chinese companies have not been granted permission to release ChatGPT-like services for public use. Baidu’s Ernie Bot and Alibaba’s Tongyi Qianwen are still in the trial phase or limited to business use only. OpenAI’s ChatGPT and Google’s Bard are inaccessible in China, with links to other foreign models also being blocked.