Stay up to date with the latest Technology regulation updates by subscribing to our free newsletter.
Every morning, you’ll receive a comprehensive email digest, called myFT Daily Digest, summarizing the most recent Technology regulation news.
The Chinese government is preparing to introduce regulations for generative AI in an effort to balance the development of this groundbreaking technology by local companies with its need for content control.
According to two sources familiar with the matter, the Cyberspace Administration of China (CAC), a powerful internet watchdog, plans to implement a licensing system that would require companies to obtain permission before releasing generative AI systems.
This requirement tightens the previously issued draft regulations in April, which allowed groups a grace period of 10 working days to register their products with Chinese authorities after launch.
The new licensing rules, part of the finalizing regulations expected to be released this month, highlight Beijing’s struggle to reconcile its ambition to lead in AI development with its longstanding censorship policies.
Matt Sheehan, a fellow at the Carnegie Endowment for International Peace, observes that this trade-off is a first for Chinese authorities who are seeking to control and censor AI-generated content while not stifling domestic companies in the race for technological advancement.
A person close to the CAC’s deliberations stated that the regulations must avoid hampering domestic tech companies, suggesting that the authorities are uncertain about their approach.
China is responding to the emergence of generative AI systems capable of rapidly creating human-like text, images, and other content based on simple prompts.
The draft rules from April state that the generated content should adhere to core socialist values and must not contain material that undermines state power, advocates for the overthrow of the socialist system, incites division, or threatens national unity.
Two other sources close to the regulators reveal that companies like Baidu and Alibaba have been in contact with regulators to ensure compliance with the rules when rolling out their generative AI applications this year.
The CAC aims to ensure that the AI technology used is reliable and controllable, with concerns about data usage, according to its director Zhuang Rongwen.
Angela Zhang, associate professor of law at the University of Hong Kong, highlights that China’s regulatory approach is mainly focused on content control.
Authorities in other countries, such as the EU, the US, and the UK, are also racing to establish regulations addressing potential abuses of AI technology.
In China, April’s draft regulations outlined specific requirements for the quality of data used to train generative AI models, including veracity, accuracy, objectivity, and diversity.
This emphasis on data quality aligns China with Europe’s approach, where regulatory scrutiny is focused on addressing issues like fabricated material generated by AI systems, known as “hallucinations”.
Matt Sheehan notes that China sets its requirements much higher, creating a challenge for Chinese companies in filtering the training data for AI systems.
However, there is a shortage of available data that meets these demands, which limits many companies in developing and improving large language models, the technology behind chatbots like OpenAI’s ChatGPT and Google’s Bard.
Helen Toner, director of strategy and foundational research grants at Georgetown University’s Center for Security and Emerging Technology, predicts that businesses will be more cautious and conservative in AI development due to the potential severe consequences of violating the regulations.
The draft regulations suggest that tech companies will bear the primary responsibility for any content created by their AI models, discouraging companies from making their models available to avoid being held accountable for issues beyond their control, adds Toner.
Attempts to contact the CAC for comment by the FT have been unsuccessful.
Additional reporting from Ryan McMorrow in Beijing
Denial of responsibility! VigourTimes is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
