Last week, Shanghai held the World Artificial Intelligence Conference (WAIC), China’s largest AI event, which attracted 500 exhibitors, 1,500 displays and more than 300,000 attendees, including Chinese Premier Li Qiang.
But despite its impressive size, the conference left me disappointed. I was hoping to witness technological advances in the field. Instead, WAIC confirmed my suspicions: there is a gap between China’s AI capabilities and the cutting-edge innovation coming out of Silicon Valley.
WAIC exhibitors focused on robotics and large-scale language models (LLMs), with only a few companies working with generative AI. More than half of the exhibitors at WAIC, including major technology companies and national telecommunications companies, showed new models.
In Shanghai, Baidu founder Robin Li encouraged attendees to start developing practical AI applications rather than continuing to refine their law masters, emphasizing that powerful, widely used AI applications will benefit society more than alternative models that can process vast amounts of data but have no practical use.
The generative AI applications on show in Shanghai were mostly chatbots like ChatGPT, with the exception of Kling, a text-to-visual application from Kuaishou that I was really impressed with, a product similar to Sora.
As I walked around the showroom, I noticed that most of the chatbots required prompts in English rather than Chinese, which led me to suspect that many of China’s AI programs are actually running on models developed outside of China.
The model clearly still needs some tweaking: One consumer ran an image of “a cute little boy with brown hair sitting in a garden” through Moore Threads’ text-to-visual app, resulting in a baby with light fuchsia skin, misaligned eyes and a disproportionately small body.
I left the conference agreeing with Alibaba Chairman Joe Tsai’s candid admission earlier this year that China’s generative AI development is at least two years behind the US, meaning US and Chinese companies aren’t really in the same league and direct comparisons are difficult to make.
A key problem is that China’s LLMs are limited to using data inside the Great Firewall. As investment bank Goldman Sachs pointed out late last year, “LLM performance improves with scale: more parameters, higher quality and quantity of training data, more training runs, and more computation.” There is simply less information on the isolated Chinese-language internet compared to an open internet with sources in a variety of languages.
AI companies outside of China have much more data to use for training, and Chinese AI developers will have a hard time keeping up with the pace.
The constraints of limited access to advanced GPUs are also evident: US policies that restrict access to cutting-edge chips and chip-making technology mean Chinese companies fall behind their non-Chinese peers.
But despite these limitations, Chinese AI developers are exploring opportunities to innovate.
Many talented people from China’s mature consumer technology ecosystem are turning to AI. Most of the founding members of the much-talked-about “Four Tigers” — Baichuan, Zhipu AI, Moonshot AI, and MiniMax — have worked in large technology companies. Their strong consumer and product intuition is why they are currently leading the AI application field in China. From a consumer perspective, their products are on par with many of the leading applications in the United States.
There are also advancements on the hardware side, with Huawei’s Ascend AI processors in particular appearing to be way ahead of the competition. The Chinese tech giant, which currently uses chips from SMIC, claims that its Ascend 910B AI chip can outperform Nvidia’s A100 chip in some tests, especially when used to train large AI models.
China’s AI developers face several fundamental obstacles, including a difficult environment, a shortage of advanced chips, geopolitical isolation, and national security concerns that limit the mobility of talent and capital.
These constraints will result in the creation of two parallel AI ecosystems, one inside China and one outside the country, with the United States likely to continue to take the lead in developing this transformative technology.
But the U.S.’s technological dominance doesn’t mean Chinese AI developers are being left behind. Chinese companies have always started out a step behind their non-Chinese peers, but fierce competition and a willingness to experiment have allowed them to catch up with the rest of the world and, in the case of consumer internet companies, get an edge over their peers.
In the world of AI, the United States and China are both friends and rivals, and we should hope that geopolitical rivalry between the two countries does not stifle innovation and cooperation.
Opinions expressed in commentary articles on Fortune.com are solely those of their authors and do not necessarily reflect the opinions or beliefs of Fortune.
