China rolls out interim measures for safe AI usage, lessens restrictions

Getting your Trinity Audio player ready...

After months of heated deliberations, the Chinese government has outlined new artificial intelligence (AI) regulations to standardize innovation in the sector.

The new rules, dubbed “Interim Measures for the Management of Generative Artificial Intelligence Services,” became law on August 15. The rules containing 24 clauses are a collaboration between eight Chinese ministries, including the Cyberspace Administration of China (CAC).

The rule book mandates generative AI content to conform to existing socialist ideals and “must not generate incitement to subvert state power.” The regulation warns developers to ensure that AI models are trained to prevent discrimination against users based on ethnicity, age, and gender and should not “promote Content prohibited by laws and administrative regulations such as ethnic hatred, ethnic discrimination, violence, obscenity, and false harmful information.”

Developers are urged to respect existing intellectual property laws by seeking the consent of IP holders before using their content to train AI models.

Article 8 of the rules require clear labeling on all AI-generated content while pushing for standardized labeling procedures.

“Providers shall legally assume the responsibility of network information content producers and fulfill network information security obligations,” a clause in the rules read. “If personal information is involved, the responsibility of the personal information processor shall be assumed in accordance with the law, and other obligations to protect personal information shall be fulfilled.”

Other provisions include the requirement of AI service providers to make necessary disclosures to relevant national authorities. AI developers are expected to provide information on the “scale, type, labeling rules, algorithm mechanism” of their training data.

Per the provisions, users of generative AI services have the right to report erring service providers for violating the rules. The rule book does not expressly stipulate a fine for violation of the clauses, opting to issue warnings and a suspension of service if the breach continues.

China is seeking the development of home-grown generative AI platforms over foreign service providers like OpenAI’s ChatGPT and Google’s (NASDAQ: GOOGL) Bard.

Experts have termed the rules as an attempt by the Chinese government to gain a tighter grip on the local AI industry, drawing similarities with its tight leash on the digital asset industry. Others have pointed to a brewing faceoff between China and the U.S. over semiconductors, leading to Chinese administrators clamping down on the export of chipmaking raw materials.

Treading the fine line of AI regulation

Global regulators are challenged to opt for the ideal path for controlling AI development. While all governments agree that the safety of consumers is of paramount concern, there is the risk that stiff regulations may hinder innovation.

The stance has been echoed by a coalition of firms led by Github, urging the European Union (EU) to soften its rules on open-source AI. There are growing whispers that AI developers are shopping for friendlier jurisdictions due to tighter U.S. and European regulations.

Watch: AI, Blockchain, and secret to winning in technology