The race for artificial intelligence (AI) and its supremacy has been brewing over the last decade and has been on the rise. Numerous nation-states and non-State actors have been vying to develop and advance their capabilities in AI. It has emerged as a benchmark and lynchpin in the projection of power and technological capabilities. China has been at the forefront of these advances and is likely amongst the major powers that hold the mantle, followed by countries like the US, Russia, Japan, India, and Europe. Amongst the various advances in AI, the emergence of software and applications like ChatGPT, Bing AI, Stable Diffusion, BloombergGPT, DALL-E and other forms of generative (Gen AI) tools have become a prominent topic of discussion. AI-generated content and the construction of a parallel Gen AI universe have raised concerns in the People's Republic of China (PRC).

Now, what are the Gen AI and the related models that are in use? According to Rouse (2023), Gen AI "is a broad label that's used to describe any type of artificial intelligence (AI) that can be used to create new text, images, video, audio, code or synthetic data". It includes the generative processes of machine learning (ML) as well as deep learning, which layers neural networks together and is improved by training these models on large amounts of data, using large amounts of computing power. These large language models are the main reason associated with the technological advances and developments in deep fakes and ChatGPT, essentially through the training received via large swathes of the entire internet in terms of text data. ChatGPT can edit an essay or edit a video, summarise books, write poetry, give advice, and provide large codes to hack critical internet infrastructure (CII) and even drones, according to the user's request. Like general AI, Gen AI includes learning algorithms for predictive processes and purposes mimicking natural ecosystems and their processes. However, the PRC's advances in the field of Gen AI have mainly been lacking, despite having multi-billion tech companies like Baidu, Tencent, iFlyTek, SenseTime and Alibaba undertaking massive research and patenting in the field.

 

Credit: news.cgtn.com

According to Dr Jeffrey Ding, Associate Professor at George Washington University, the Gen AI chatbot technology in the PRC has lagged behind the US and other major technology powers. The Chinese tech company Baidu launched its own AI- language model chatbot named Ernie Bot, while Alibaba unveiled the Tongyi Qianwen, iFlyTek launched SparkDesk. However, despite the investments, these are still considered lagging compared to ChatGPT. Baidu lost one-tenth of its stock value in the aftermath of the presentation by the Chairman of Baidu unveiling the Gen AI technology. Dai Qionghai, a computer scientist based at Tsinghua University who is part of the Chinese Association of Artificial Intelligence, argues that China is "performing at a high level and getting close to the level of the US". However, as a major power concerned with the use and utilisation of Gen AI in its society, the PRC has been undertaking various steps to deal with Gen AI.

Various challenges perceived by the PRC are emerging due to the widespread use of Gen AI tools for the country. This includes the proliferation of what is characterised and categorised as false information, disinformation and propaganda inimical to the PRC and its functioning; issues related to security, privacy and legitimate rights as well as interests of Chinese citizens; the risk of creating discriminatory and inflammatory tensions within the Chinese society as well as with regard to public opinion; and finally issues and problems related to social ethics and related security challenges as well as risks. Furthermore, there have been many discussions about the threats of Gen AI, including but not limited to false, discriminatory and biased information, data privacy issues, public safety, intellectual property rights, governance, ownership, and other issues. The essence of these threats can be traced back to the data sources and algorithm models.

In the PRC, the China Internet Network Information Center (CNNIC) (Zhongguo Hulian Wangluo Xinxi Zhongxin), in conjunction with telecommunications, public security, market supervision and other departments heads the section related to Gen AI. On 11 April 2023, the Cyberspace Administration of China (CAC) released the "Measures for Generative Artificial Intelligence Services (Draft for Comment)" (hereafter referred to as Measures), aimed at the promotion of the healthy development and standardised application of Gen AI. The Measures pertain to an important legislation, which is a response of national regulators to the new Gen AI technologies such as ChatGPT, which has recently exploded and spread across the global market. The provisions of Gen AI services in the PRC have entailed requirements and institutionalising regulations that adhere to the norms of social morality, public order, and customs in the country. The Measures sets up forwarding clauses, supervision, inspection, and legal responsibilities into three fundamental laws in cybersecurity, network security and data privacy protection in the PRC. This includes the "Cybersecurity Law of the People's Republic of China, 2017", "Data Security Law of the People's Republic of China, 2021", and "Personal Information Protection Law of the People's Republic of China, 2021".

Furthermore, the PRC is encouraging indigenous innovation (Zizhu Chuangxin) (mentioned in Art 3 of the Measures), which is also a central strategic concept of Xi Jinping's emphasis on self-reliance (ziliziqiang) in industry and technology as well as in the 14 th Five-Year Plan of the PRC released in March 2021. In future, when Gen AI is combined with terminal products, such as for cultural entertainment, medical treatment, smart cars, etc., regulation, surveillance and censorship are expected to increase depending upon the perceived threat and vulnerability level. However, in terms of regulatory focus, the consultation draft is only applicable to products in the Chinese public domain and does not impose too many restrictions on Gen AI tools and applications that are still in the prototype/laboratory training stage and non- public trial operation. The regulatory objects of the Measures are mainly organisations and individuals that use Gen AI products to provide services such as chat and text, image, and sound generation, including supporting others to generate text, images, and sounds by themselves by providing programmable interfaces. However, even if there are fewer legal restrictions on Gen AI products in the prototype/laboratory training stage and non-public trial operation stage, developers still need to abide by laws and regulations related to network security and data privacy protection and should provide for the products after they are launched on the market, and make adequate plans for compliance as well.

In relation to Gen AI technology and applications, there are two previously promulgated departmental regulations, viz., the "Internet Information Service Algorithm Recommendation Management Regulations" and "Internet Information Service Deep Synthesis Management Regulations". The "Internet Information Service Algorithm Recommendation Management Regulations" mentions the algorithm recommendation technology of generating and synthesising and requires the information generated and synthesised by the algorithm to be marked clearly. Those engaged in Internet news information services are prohibited to generate synthetic false news information. The "Regulations on the Administration of Deep Synthesis of Internet Information Services" stipulates in more detail the requirements that deep synthesis services that have the function of generating or significantly changing information content, including intelligent dialogue, intelligent writing, face generation, face manipulation, and gesture manipulation, should comply with. The public is reminded in the form of prominent signs, establishing and improving management systems such as user registration, algorithm mechanism review, scientific and technological ethics review, information release review, data security, personal information protection, anti-telecom network fraud, emergency response, etc., and strengthening training data management, technical management, etc.

The regulatory landscape in the PRC, especially with the emergence and institutionalisation of the Measures, along with pre-existing regulations and legislations, is expected to increase compliance, surveillance and censorship in Chinese society. However, this will have a huge impact on the (indigenous) innovation tendencies within the ambit of the Gen AI technology ecosystem in China and amongst major Chinese tech companies, which are already lagging in comparison to their counterparts in the US and Europe and thus expected to increase the gulf between Chinese tech companies and those in the West. This, along with an increasing decoupling tendency between China and the US, will have a huge impact on the manner in which Gen AI technology can be developed by Chinese companies, who have historically been dependent on technology transfer, reverse engineering and on (purportedly) (il)legal exchanges of intellectual property (IPR) with the West. However, the recent trajectory of initiatives and activities by China is a much-needed addendum in a new era in which disinformation and post- truth tendencies have been on the rise, wherein several complex and transformative challenges have emerged to nation-states across the world.

(Author is Assistant Professor of Politics, Department of International Studies, Political Science and History, CHRIST (Deemed to be University), Bengaluru, Karnataka, India, Email: [email protected]. Views expressed in the article are personal to the author, and do not necessarily reflect the views of AICIS.)

References

1. China Internet Network Information Center (CNNIC). (2023). https://www.cnnic.com.cn/

2. Creemers, R. & Webster, G. (2021). Translation: Personal Information Protection Law of the People's Republic of China – Effective Nov.1, 2021. https://digichina.stanford.edu/work/translation-personal-information-protection- law-of-the-peoples-republic-of-china-effective-nov-1-2021/

3. Creemers, R., Webster, G., Triolo, P. (2018). Translation: Cybersecurity Law of the People's Republic of China (Effective June 1, 2017). Stanford University. https://digichina.stanford.edu/work/translation-cybersecurity-law-of-the-peoples- republic-of-china-effective-june-1-2017/

4. Dai, K. and Deng, J. (2022). China's Personal Information Protection Law (PIPL). Bloomberg Law. https://pro.bloomberglaw.com/brief/china-personal-information- protection-law-pipl-faqs/

5. Huang, S., Toner, H., Haluza, Z., Creemers, R., & Webster, G. (2023). Translation: Measures for the Management of Generative Artificial Intelligence Services (Draft for Comment) – April 2023. Stanford University. https://digichina.stanford.edu/work/translation-measures-for-the-management-of- generative-artificial-intelligence-services-draft-for-comment-april-2023/

6. Inkster, N. (2020). The great decoupling: China, America and the struggle for technological supremacy. London: Hurst.

7. National People's Congress of the People's Republic of China (NPC). (2021). Data Security Law of the People's Republic of China – Order of the President of the People’s Republic of China. http://www.npc.gov.cn/englishnpc/c23934/202112/1abd8829788946ecab270e469 b13c39c.shtml

8. Ning, S.X. and Wu, H. (2023). Cybersecurity Laws and Regulations China. https://iclg.com/practice-areas/cybersecurity-laws-and-regulations/china

9. Waldie, B. (2022). Lexicon: Indigenous Innovation or Independent Innovation (Zìzh? Chuàngx?n), Stanford University. https://digichina.stanford.edu/work/lexicon-indigenous-innovation-or-independent- innovation-zizhu-chuangxin/