How China is betting cheap AI will get the world hooked on its tech

Artificial intelligence (AI) is at a very Chinese time in its life. Recent moves from Chinese AI labs are throwing the dominance of American “frontier labs” such as Google and OpenAI into question.

Last week ByteDance, the company behind TikTok, released an AI video-generating tool called Seedance 2.0 which produces high-quality film-like clips from text prompts, with a casual disregard for copyright concerns. This week Anthropic, the US company behind the chatbot Claude, said three Chinese AI labs created thousands of fake accounts to harvest Claude’s answers in a practice called “distillation” which can be used to improve AI models.

These events have led to suggestions that China may be gaining the upper hand in the battle to dominate AI. So, is China winning the “AI race”?

Cheap, widely used tools

While most advanced frontier models are still made by American companies, China is pushing hard to develop cheap, widely used AI tools, which could create global dependence on Chinese platforms.

Reuters reports the industry is bracing for a “flurry” of low-cost Chinese AI models, with Chinese systems repeatedly driving usage costs down.

What’s the plan? China’s official AI policy documents suggest China sees AI as “a new engine for building China into both a manufacturing and cyber superpower”, and “a new engine of economic development”.

Since 2017, China has recognised that the technology is at the centre of “international competition”. “By 2030,” one key policy document says, China’s AI “technology and application should achieve world-leading levels, making China the world’s primary AI innovation center”.

This focus on becoming the dominant player in AI helps explain why Chinese firms are pushing hard on price. If you can make your AI cheap enough, you might just make it globally ubiquitous.

Cost helps determine who adopts AI first, and which models are first implemented in software and services. Even if the United States remains ahead on most elite benchmarks, Chinese products could still become globally influential if they are widely used and widely depended upon.

High-tech soft power

But China does not present its AI technology to the world as only benefiting itself. Instead, it’s pitched as a contribution to humanity.

A 2019 statement of “governance principles” from a national AI governance expert committee argues that AI development should enhance “the common well-being of humanity” and “serve the progress of human civilization”.

These phrases portray AI as a technology that advances the human story itself, rather than only serving Chinese interests. It suggests Chinese AI leadership is good for everyone.

This is an example of Chinese soft power. Tools such as Seedance may threaten Hollywood’s business model, but they do something else too. High-quality, low-cost generative media can spread quickly.

If Chinese systems become widespread, they can influence creators, developer habits, and platform dependencies, especially in non-Western markets that need affordable tools and may dislike American tech dominance.

The spread of the ‘Chinese model’

For liberal democracies such as the United Kingdom, Australia and Canada, the growth of Chinese AI tools creates a strategic headache. It will not be easy to manage security concerns about Chinese technology while avoiding technological isolation if Chinese AI tools become widely adopted.

There is a darker side to China’s AI tools. US think-tank Freedom House describes China as having the world’s “worst conditions for internet freedom”, and suggests other nations are now “embracing the ‘Chinese model’ of extensive censorship and automated surveillance”.

In 2022, the Cyberspace Administration of China issued rules for the algorithms that curate news feeds and short video platforms. Providers are required to “uphold mainstream value orientations” and “vigorously disseminate positive energy”.

These algorithms are important because they shape what people see and what is suppressed. As a result, these rules suggest the Chinese government is deeply concerned with controlling information across its social media platforms and AI tools.

A dilemma for third parties

Not every Chinese AI tool is a propaganda weapon. Rather, China is building world-class AI technology within an authoritarian system that prioritises the control of information.

This means China’s ability to make generative AI commercially powerful will likely also, despite its claims about serving “human civilisation”, make censorship and narrative management cheaper and easier.

China’s business and soft-power model is a much bigger story than just Seedance’s cavalier attitude towards copyright or Anthropic’s concerns about intellectual property. China’s goal is to build AI tools that rival those created by America’s tech giants, and to make them inexpensive and adopted globally.

For other countries, this may create a dilemma. Once a technology becomes a standard, it can be difficult to justify using a different product.

The question that remains is whether liberal democracies can adopt China’s low-cost products without drifting into dependence on systems shaped by an authoritarian political model.