Tech

China-U.S. AI Arms Race Heats Up as Chinese Startup Unveils Powerful New AI

A New Chinese Startup’s Powerful AI Is Already Beating Meta’s Llama 2, Analysis Finds

A new Chinese artificial intelligence startup, spearheaded by one of the field’s foremost experts, catapulted itself onto the world stage this week when it released an open-source AI model that appears to outperform Meta’s own model in critical metrics, giving the country an early win in what the company’s founder sees as a battle between the U.S. and China for AI supremacy. 

The company, called 01.AI, is the brainchild of CEO Kai-Fu Lee, the former head of Google in China and co-chair of the World Economic Forum’s Artificial Intelligence Council, who authored the bestseller AI Superpowers: China, Silicon Valley and the New World Order, a 2018 book which argues that China is capable of overtaking the U.S. to be the world leader in AI technology. Lee reportedly only began assembling his team in March but has quickly hired more than 100 employees who come from both multinational companies and Chinese firms. 01.AI, which began operating in June, has already been valued at over $1 billion by its investors, which include Lee’s own venture firm, Sinovation Ventures, and Alibaba Cloud. 

Videos by VICE

On Sunday, 01.AI stormed out of the gate when it released Yi-34B, an open-source large language model, or LLM, that Lee described this week as a “compelling alternative” to Meta’s LLaMA 2. LLMs are AI models similar to ChatGPT that are trained on a large amount of data to generate text, although they can be applied to other tasks. The model, which is available in English and Chinese, has quickly garnered attention for scoring first among for scoring first over the weekend on the open-source community Hugging Face’s rankings of so-called pre-trained base LLMs, outperforming even Meta’s celebrated model. 

In an interview with Bloomberg, Lee said LLaMA 2 “has been the gold standard” for the open-source community, but that he wants “to provide a superior alternative not just for China but for the global market,” reiterating on Twitter Monday that his “underlying vision” was to “make better AI accessible to more people.” He expects his globally available model will prove useful to multinational banks and insurers, among others, Lee told Bloomberg.

Lee told TechCrunch this weekend that while China is still behind the U.S. on LLM technology, there is “no doubt China can build better applications than American developers” because of the country’s “phenomenal mobile internet ecosystem.” That appears to be where the company will put its focus. “The duty is not just to push out good research models, but even more importantly to make application development easy so that there can be compelling applications,” he said. “At the end of the day. It is an ecosystem play.”

In anticipation of the U.S. further restricting Chinese access to the chips necessary to build out AI models, 01.AI reportedly started stockpiling them earlier this year, borrowing money from Lee’s venture firm in order to do so. “We basically bet the farm and overspent our original bank account,” he told Bloomberg. “We felt we had to do this.” Lee told TechCrunch in a separate interview that his company now has enough chips to last at least a year. 

In interviews, Lee has positioned 01.AI as a necessary response to U.S. restrictions that have curtailed China’s ability to advance in the AI field. “Unlike the rest of the world, China doesn’t have access to OpenAI and Google because those two companies did not make their products available in China, so I think many doing LLM are trying to do their part in creating a solution for a market that really needs this,” Lee told TechCrunch. 

The stakes are clear to Lee, who has written extensively about what he sees as the coming battle between China and the U.S. for AI supremacy. 

Born in Taiwan, Lee moved to the U.S. during high school and went on to major in computer science at Columbia University. Since then, he has largely dedicated his life to the AI field. According to Bloomberg, he stated in his 1982 graduate school application to Carnegie Mellon that he planned to dedicate his life to the field. For his doctoral dissertation six years later, Lee created what he claimed was “the world’s first large-vocabulary speech-recognition model,” according to Time. He went on to serve as the head of Google in China, as well as a top executive at both Microsoft and Apple, before starting his own venture firm.

But it is in his writing on the subject of AI that Lee has become best known. He has penned two best sellers on the subject,  AI Superpowers and AI 2041: Ten Visions for Our Future, and repeatedly warned of the potential economic and social upheaval that he expects AI will bring about before the end of the century.

In a 2017 opinion piece for The New York Times, Lee suggested the AI technology would lead to even more wealth concentration, rising profits for corporations, and mass unemployment. This development would lead to the “unavoidable” decision to raise taxes on corporations who profit from AI and transfer “large chunks” of the money to “those whose jobs have been displaced,” he argued. 

Internationally, Lee envisioned a world in which the U.S. and China become the two dominant players in the AI space, and all other countries must fall in line, exchanging international welfare for the parent country’s ability to profit off their local population. “Unless they wish to plunge their people into poverty, they will be forced to negotiate with whichever country supplies most of their A.I. software—China or the United States—to essentially become that country’s economic dependent, taking in welfare subsidies in exchange for letting the ‘parent’ nation’s A.I. companies continue to profit from the dependent country’s users,” he wrote. 

This year, Time named Lee one of the most critical people in the field. In an interview, he said more regulation is needed to prepare for the change that is about to occur. 

“A lot more needs to be done,” he said. “When AI is this powerful, able to come up with things that we didn’t know before, it might be used to come up with new ways of harming others, of creating weapons, [or] using misinformation to manipulate people for profit or evil intent.”