Australian entrepreneurs and investors are already looking to new developments in our region for inspiration and opportunities. They probably don’t look much more than China.
China is creating an ecosystem attractive to global artificial intelligence (AI) talent and capital. A US-led tech lash against Chinese companies and suspicion of scientists of Chinese background has been used by China to attract talent to do research and business. This certainly suggests that nurturing and attracting AI talent and protecting copyright could be areas for Australia to invest in while developing its own AI capabilities.
Industry insiders believe that China’s market size is its advantage in the big data industry. China’s rapid integration of AI into existing products and services, widespread adoption of intelligent lifestyles, and flexible approach to user privacy and data protection for domestic companies and research institutions provide more opportunities for new ideas. In addition to an ecosystem conducive to AI start-ups and research, China’s approach to developing and regulating generative AI could attract global entrepreneurs and businesses looking for alternatives to Euro-American models.
China’s response to ChatGPT is a new moment in the AI “Great Leap Forward”. A fierce race by its big tech players to develop generative AI and foundation models is known as the “Battle of One Hundred Models” (baimo dazhan). It is worth watching how this battle pans out and holds some lessons for Australia.
The AI Great Leap Forward has gained momentum since China’s “Sputnik Moment” in 2017 when AlphaGo claimed victory over China’s best Go player. In the same year, China’s State Council announced a national plan to become a world leader in AI by 2030 and allocated $59 billion in funding for R&D in AI software, hardware, intelligent robotics and vehicles, virtual reality, and augmented reality. AI was seen as a potential competitive advantage in manufacturing capacity (as in the “Made in China 2025” plan) and as an opportunity to take on a larger rival such as the US. Since then, Chinese technology companies have been bringing generative AI models to application fields such as media, arts, education and entertainment. They are also investing in global start-ups specializing in AI critical technologies.
China’s approach to generative AI is ambitious but cautious. Like many countries, China has ethical and regulatory concerns. Developing an accountable and secure framework, prioritizing research and training skilled workers to meet new technological and regulatory challenges is considered as important as promoting domestic innovation and intellectual property protection. Already an AI superpower with a projected AI economy of USD7 trillion by 2030, China will inevitably play a significant role in global AI governance and standard setting, according to a PricewaterhouseCoopers study.
Western observers have noted the Chinese Communist Party’s (CCP) obsession with information control. For example: Nicholas Welch and Jordan Schneider note that Chinese regulators require AI-generated content to “reflect core socialist values and … not contain content that violates state power.” They point to a potential collision between the country’s AI aspirations and its censorship regime.
Industry and academic observers in China have a different view. They emphasize the balance between content control and innovation/creativity in the development, rollout, and regulation of large language models (LLMs) to win the AI race. The battle of a hundred models signals intense competition among Chinese technology companies in the expensive and protracted LLM race. The race is led by big players such as Baidu, Alibaba, Tencent, Huawei and SenseTime, followed by medium and small players that focus on narrow or vertical models that are faster to train, develop and apply across industries. General AI or AGI (Artificial General Intelligence).
China has made significant progress in AI applications, such as face and voice recognition, autonomous vehicles, robotics, and AIGC (AI-Generated Content). This technology is widely used in automatic driving, e-commerce, healthcare, social media platforms, media and art. AIGC and the vertical LLM aim to unlock China’s software-as-a-service (SaaS) potential and increase corporate productivity.
The global publicity and widespread adoption of generative AI has led to controversy regarding privacy and data security. DeepFake is an example of facial recognition technology that can be used for nefarious purposes. In China, a face-swapping mobile app called ZAO was quickly ordered off the shelves in 2019 for violating user privacy. In January 2022, the Cyberspace Administration of China (CAC) issued algorithms and “deep synthesis tech” regulations, which will apply from January 2023, to ChatGPT-style technologies.
In April 2023, the CAC proposed new regulations governing generative AI in China, focusing on the AIGC. Generative AI places this significant responsibility on service providers, including data and content labeling and filtering to ensure accuracy, respect for intellectual property, and avoidance of bias. In July the CAC issued “Interim Measures for the Management of Generative Artificial Intelligence Services”. Unsurprisingly, the emphasis is on balancing AI’s economic dynamism and state control, that is, strategically using AI to advance China’s commercial, technological, and political goals while creating an adaptive regulatory framework in response to new cases and application scenarios.
China’s flexible and iterative approach to regulation is well suited to adapt to rapidly changing technology. China scholar Roger Cremers has pointed out that these regulations can lead companies to prioritize services and products in industrial and public sector applications.
The Carnegie Endowment for International Peace characterized this approach to AI regulation as “vertical” (targeting specific AI applications or sets of applications), which contrasts with the European Union’s horizontal approach (targeting most applications of the technology). China also takes a forward-thinking approach to regulating the input and output of LLM, paying attention to the social impact of generative models.
Australia has many comparative advantages in digital innovation and multiculturalism. The Tech Council of Australia notes that we can capture the potential of generative AI and strengthen our AI capabilities through global collaboration, particularly with our regional players. Australia can learn from China to invest in the research, computing and infrastructure needed to develop generative AI and its applications. We can leverage our geographical and geopolitical advantages and multicultural human expertise by taking an active role in regional and global dialogues on AI regulation, AI innovation and their trade-off effects on society.
Read more articles in our China Perspective series:
China: Perspectives beyond the mainstream media