China’s AI technology is lagging behind that of the U.S., experts say.
AFP/Getty ImagesTechnology has long been central to the increasingly heated race for global military and economic supremacy between China and the U.S. But the rapid advancement of artificial intelligence adds another complicated layer to the two countries’ already fraught relationship.
While China appears to lag behind the U.S. in AI technology—for now—experts say U.S. lawmakers must keep the nation in mind when deciding the extent of AI regulation. That arms race—along with China’s own heightened regulation and some high-profile AI stumbles—is leaving investors on the sidelines until the technology matures in the world’s second-largest economy.
AI’s complex nature and meteoric development has bred confusion in both countries among government officials, investors, and even companies trying to get into the AI arena. U.S. lawmakers have apparently been so flummoxed by the emerging technology, they have asked AI firms themselves what regulations are needed.
There is also widespread disagreement over how much Washington could hobble China’s programs with a more severe ban on the semiconductors that are crucial for AI. Last year, the U.S. banned sales to China of an industry-leading Nvidia (ticker: NVDA) chip central to many AI programs.
Nvidia later began to sell a slightly slower chip to China that isn’t covered by the ban. Experts say the move won’t meaningfully handicap AI programs in China. Meanwhile, Baidu (BIDU) has developed its own chip to reduce China’s import dependency.
What this all means for Washington’s strategy to contain China’s tech growth has been a particularly messy question. Influential industry experts—including the CEO of industry leader OpenAI—have argued that regulating AI could stall its development in the U.S. and give China a window to leap ahead.
But this view is misleading at best and dangerous at worst, Georgetown AI expert Helen Toner and colleagues said in a report this week. China has been rolling out AI rules earlier and more frequently than the U.S., and its most recently proposed regulations are “pretty stringent,” she said.
Moreover, a closer look at China’s AI prowess reveals more bluster than substance compared with the U.S. Toner’s team spoke with Chinese AI researchers who said China’s large language models (LLMs)—akin to OpenAI’s ChatGPT— “are at least two or three years behind their state-of-the-art counterparts in the United States—perhaps even worse.”
This inflated sense of China’s AI sophistication is understandable. As in the West, there are near-daily reports of industry-changing breakthroughs in various Chinese AI projects. And the list of firms that have joined the AI bonanza there is long, with some creating dedicated departments or integrating the technology into existing product lines.
Firms entering the fray include tech heavyweights like Baidu, Alibaba Group Holding (BABA), Huawei Technologies, Tencent Holdings (700), and NetEase (NTES), to name a few. Chinese universities and the government are well in the game, too.
But there have been little impressive results so far. Baidu’s much anticipated ChatGPT-style “Ernie Bot”—perhaps the most famous AI project in China—malfunctioned during its unveiling in March, and has left reviewers unenthused since. Several other Chinese firms’ new tools have gotten lackluster receptions as well.
Rollouts haven’t been perfect for U.S. counterparts, either—Google’s Bard chatbot made a factual error during its first public demo. But the response to ChatGPT and its U.S. peers has been generally positive, and, within the stock market, particularly feverish.
Also tracking AI progress in both countries, of course, are investors. The main questions are how overhyped are China’s AI programs could be and how Beijing’s longtime practice of tightly controlling information could affect their growth.
Newsletter Sign-up
A successful AI investment strategy in China, analysts say, requires taking a cue from the government itself. In other words, investors should consider “aligning with the government’s objectives and taking advantage of available incentives,” says Asia-focused business intelligence firm Dezan Shira & Associates.
Though Beijing has rolled out extensive regulation of AI, it has also said it will ensure the industry’s growth will benefit areas it deems geopolitically and economically strategic, such as aerospace, supercomputing, robotics, and telecommunications, among other fields.
“China will focus strongly on AI technologies it sees as crucial for national security, which by now has come to mean a lot more than it did even five years ago,” Rogier Creemers, a professor at the University of Leiden who specializes in Chinese digital technology policy, told Barron’s.
Another sweet spot for investors could be sectors where censorship is of low concern to Chinese officials, and where the benefits of AI could be significant, such as self-driving cars and healthcare.
“While many things in China are covered by censorship, many are not—and these are capable of quite some [AI] application,” Creemers added.
UBS analysts, in a recent report, singled out areas they expect to benefit from AI’s rise in China, including e-commerce, online ads, gaming, and molecular biology, if the country can continue to develop better domestic computer components.
Still, AI-related stocks in China remain pricey, largely driven higher by an unfettered retail investor buying spree rather than stock analyst recommendations.
“The development of generative AI systems is in a nascent stage,” Jia Tan, head of China equity research at UBS, wrote in the report. “While potential applications across a wide range of industries seem limitless, runaway stock valuations of these Chinese companies are not supported by fundamentals.”
That—along with a murky regulatory environment, technology hiccups, and U.S. tensions—gives investors plenty of reasons to stay on the sidelines and see how China’s AI industry plays out.
Write to editors@barrons.com
