China’s new generative AI regulation sets guidelines to promote values of ‘socialism’

The Cyberspace Administration of China (CAC), the nation’s lead internet regulator, has unveiled the country’s inaugural rules overseeing generative artificial intelligence. Lauded as a significant regulatory stride, these rules, which take effect on Aug. 15th, were established when Beijing heightened its vigilance over burgeoning technology.

These regulations mandate generative AI services to uphold the “core values of socialism,” underscoring China’s distinct socio-political backdrop of AI development compared to the U.S.

The Global Boom in Generative AI.

Generative AI, which can generate text or images, is witnessing a global boom. Prominent U.S. entities, including OpenAI’s ChatGPT, have been at the forefront of this wave, pioneering a surge of comparable services. CNBC reports that these technologies’ swift and widespread proliferation has raised worldwide regulatory concerns.

China is no outlier in this trend, with CNBC noting that tech behemoths such as Alibaba and Baidu are actively refining their generative AI services. Yet, the country’s uncompromising control over the domestic internet and its monitoring of AI advancements have instilled a careful strategy among these entities. As Reuters notes, Beijing’s increasingly rigorous tech regulations have slowed the public release of generative AI services, many of which are still beta testing.

In the face of such challenges, the CAC’s rules aim to facilitate the responsible development and usage of generative AI. As per the new regulations, generative AI services, notably those accessible to the public, will mandate an operating license. CNBC reports that service providers encountering “illegal” content must cease its creation, refine the algorithm, and report the content to the authorities. Furthermore, they must carry out security assessments and safeguard user data.

China’s generative AI regulation provisions.

According to the CAC website, the new regulations’ core aim is to promote the healthy development and standardized application of generative AI, safeguard national security and public social interests, and protect the rights and interests of citizens, legal entities, and other organizations.

The CAC explicitly set out the following detailed provisions for the new regulatory mandate:

• Adherence to the core values of socialism, no generation of content that incites harm to the state or promotes prohibited content such as violence, obscenity, or false harmful information.
• Prevention of discrimination during the AI algorithm design, data selection, model generation, optimization, and service provision, based on characteristics such as ethnicity, belief, country, region, gender, age, occupation, health, etc.
• Respect for intellectual property rights and business ethics, maintenance of business secrets, and avoidance of using algorithms, data, or platforms to implement monopoly or unfair competition.
• Respect for others’ rights, ensuring the AI does not endanger physical and mental health or infringe on individual rights such as portrait, reputation, honor, privacy, and personal information rights.
• Implementing effective measures to improve the transparency of generative AI services and enhance the accuracy and reliability of generated content based on the service type.

Despite their apparent rigidity, these rules exemplify China’s delicate balancing act in fostering AI innovation while maintaining meticulous oversight of its progress. The CAC has stated that the regulations aim to “encourage innovative applications of generative AI and support the development of related infrastructure like semiconductors.”

Critique of the new regulatory mandate.

Yet, this seemingly progressive regulatory framework has attracted some criticism. Some experts, including Henry Gao, a law professor at Singapore Management University, told Reuters that China’s stringent internet laws could potentially inhibit the country’s ability to compete with its Western counterparts in the tech sphere. Gao argues that China’s ‘preemptive regulation’ approach could “stifle the innovation and slow down the ability of Chinese firms to catch up.”

Nonetheless, Reuters states that Chinese firms are pressing forward, primarily focusing on industrial applications of AI. This strategy aligns with Chinese President Xi Jinping’s emphasis on more “hard” tech breakthroughs aimed at reducing China’s reliance on Western technology, a vital feature of the new regulatory landscape.

The CAC stated that at a time when the swift development of generative AI technology has brought forth new opportunities, as well as challenges such as the spread of false information, personal information rights infringement, data security issues, and prejudice, the measures seek to balance the growth of generative AI and safety and are designed to prevent the risks of generative AI services.

Specifics of the new mandate.

The new measures propose combining development with security, promoting innovation and law-abiding governance, encouraging the creation and growth of generative AI, and implementing cautious supervision of generative AI services.

The new specifications for generative AI services also require clarifications of data processing and labeling requirements and mandating providers to take effective measures to prevent minors from excessive reliance on generative AI services and promptly deal with illegal content. Additionally, systems for security assessment, algorithm filing, and complaint reporting are stipulated, alongside clear legal responsibilities.

China’s State Internet Information Office further emphasized that the development and governance of generative AI services require participation from the government, businesses, society, and internet users to jointly promote the healthy development of generative AI and ensure the technology benefits the people.

As the rapid evolution of generative AI continues, the pioneering regulations of China provide invaluable insights for other nations wrestling with the burgeoning technology’s oversight. Whether these rules will ultimately achieve the ideal balance between innovation and control remains to be seen. However, one thing is clear – the era of unchecked AI development is progressively fading into history.