Since ChatGPT made its grand debut last year, all attention has been on the American tech giants. But China is also carving itself a place at the top. A United Nations report released on Wednesday revealed that Chinese companies and academic institutions own the largest number of patents for generative artificial intelligence, or GenAI, from the last decade – six times more than the United States, which placed second.
When we hear GenAI, most of us think of a friendly all-knowing chatbot, but the technology can be fashioned to serve many other purposes, from creating images and videos to automating tasks and discovering new drugs to training self-driving cars.
The race to innovate has already begun. In the last decade, some 54,000 patents for GenAI programmes have been recorded globally, with a quarter of those applications filed in the last year alone, according to the World Intellectual Property Organization (Wipo).
“The patent data suggests this is an area that's going to have a profound impact across many different industrial sectors going forward,” said Christopher Harrison, patent analytics manager at the United Nations IP agency.
At the helm of that acceleration is China. Not only does it make up for nearly 70 per cent of patents, with Tencent, Ping An Insurance and Baidu leading globally, but it has published the highest number of scientific articles on GenAI between 2010 and 2023. The US follows closely in second place.
China’s approach has been to let new emerging initiatives flood its internal market. “In AI, China follows the Confucius maxima ‘Let a hundred flowers bloom, let a hundred schools of thought contend’,” Jovan Kurbalija, executive director of DiploFoundation, said. “Although this maximum had a mixed historical record as it was misused by Mao during the Cultural Revolution, it served Chinese technological growth very well for the last few decades.
A Tencent executive once described the proliferation of genAI initiatives as the “war of a hundred models”. Kurbalija suspects that the likely scenario is the ones to remain standing will be a handful of powerful companies that will have absorbed the smaller competition, like in the United States.
But quantity doesn’t mean quality. Companies can file patents to protect future revenues, according to the report. “There are so many other parameters which influence the development of AI, such as integration for health, education, and market. That's the game, and I think the United States is definitely ahead when it comes to integration,” said Kurbalija.
OpenAI ranks low in the report as it didn’t begin filing patents until 2023 and only has a handful up until now. ChatGPT is among the fastest-growing services, attracting nearly 600 million monthly visits.
Wipo’s director general, Daren Tang, said that while the patents don’t necessarily mean they will become commercially available products and services, “it gives us an insight as to what's happening upstream, then we can make more educated guesses as to what's going to happen downstream in the years to come”.
When it comes to large language models like ChatGPT or Google’s Gemini, China seems to be lagging behind. Chinese AI chatbots, like Baidu’s Ernie rolled out last year, have one problem in common: censorship. The companies have to abide by restrictions on what the programmes can say and what they can use to train them so as not to contradict the Chinese Communist Party.
But Beijing might not be too worried about that. “China prefers to use AI to grow its economy and not necessarily for societal use,” said Kurbalija. It has the advantage of serving its economic objectives – and is also less controversial. One such area is China’s booming “internet of things” market. That means embedding AI software into ordinary objects, from refrigerators to thermostats to phones or cars.
“China is the manufacturer of the world and whatever it manufactures will be increasingly augmented, supported and enhanced by AI,” said Kurbalija.
Tang highlighted this as one of the new frontiers: “Industrial innovation is beginning to blur and blend with digital innovation. Just think about a car, right? Is a car a machine, or is it a very sophisticated computer on four wheels?”
With a population of 1.4 billion people, that means large amounts of user data that can be scraped and used to further advance the technology. But it also poses security issues for the Chinese and for consumers elsewhere.
The US ban on AI microchip exports to China and OpenAI’s recent announcement that ChatGPT will no longer be available in China highlight the tensions between the two rival powers over leadership and the security implications of AI. China also has the highest number of patents of GenAI models for military purposes (82), followed by the United States (28).
As more clashes emerge, calls for global regulations may grow louder. China, the United States and the European Union were among the first to roll out regulations for AI, but it is still in early stages. “Many countries are still trying to figure out what their domestic regulatory response is and UN agencies move only at a speed in which member states allow it to move. So…I don't see WIPO being able to push into setting norms and standards in a very big way,” said Tang.
But there is a hint that even the US and China are aware that at least some common understanding is needed in the AI race. The two held a meeting earlier this year in Geneva over the risks of AI, though the details were not disclosed. On Monday, the UN General Assembly in New York adopted a Chinese-sponsored resolution backed by the US, calling on wealthy countries to support developing nations in accessing AI.
Geneva Solutions content is licensed under Creative Commons BY 4.0.
Comments
No comments on this item Please log in to comment by clicking here