Reference News Network, December 12 report: On December 10, the U.S. Consumer News and Business Channel website published an article titled "The United States' Advantage in AI Talent is Fading - Because China is Scaling Up Training for the Next Generation of Researchers," authored by Zheng Yilun. The article is excerpted as follows:

The author of "The Chip War," Chris Miller, warned several years ago that the balance of power today depends on the semiconductor supply chain across geopolitical fault lines. Now, this Tufts University historian has raised a new concern: the U.S. may lose its advantage over China in the field of artificial intelligence (AI) talent.

Miller recently told a panel at the U.S. Senate Committee on Foreign Relations that, in terms of talent, "the U.S. advantage is dangerously weakening." He pointed out that this advantage "is much more fragile than the U.S. advantage in AI chips."

Researchers from the Carnegie Foundation have also expressed the same concerns, noting that in recent years, China has trained more top AI researchers, while fewer such talents are going to the U.S.

A part of this difference lies in scale itself, especially with China's increasing educational level.

China's population is four times that of the U.S., and the number of graduates in science, technology, engineering, and mathematics (STEM) is also four times. The growth rate of higher education graduates in China is astonishing, increasing several times over just one generation. While the U.S. started much higher, the growth rate during the same period was slower, rising from 8.7% to 14.1%.

This surge in talent is reshaping corporate recruitment channels. He Xiaopeng, CEO of Xpeng Motors, told reporters that most of the company's new employees are fresh graduates.

Beijing is striving to maintain this momentum to achieve technological self-reliance. The Ministry of Education stated that in the past two years, one-fifth of the national higher education courses were reformed, with a series of majors canceled or added, aiming to bring more students into fields such as artificial intelligence and integrated circuits.

In an email to me, Miller said that this does not mean China is doing better in attracting global AI talent, but being able to retain more AI experts "could have a significant impact on talent mobility."

At the same time, he said, U.S. immigration policies may make it difficult for AI researchers to work in the U.S.

The U.S. Open Artificial Intelligence Research Center (OpenAI) sparked a wave of generative AI with ChatGPT-3.5 in 2022.

DeepSeek, a Chinese company, made headlines earlier this year, claiming to have surpassed OpenAI's achievements at an extremely low cost.

Despite these restrictions, Chinese companies are increasingly using their scale advantage to gradually build domestic AI capabilities, whether in terms of engineers, chips, or data. This can be seen from the fact that Chinese AI models can match OpenAI at extremely low costs.

Chinese developers are also achieving more at lower costs.

Xiong Wei, an analyst at UBS Securities covering China's internet sector, noted in a recent report that China's internet giants invested about 400 billion yuan in capital expenditures this year, which is about one-tenth of what their U.S. counterparts invested, yet achieved "almost" the same AI model performance.

Analysts sometimes warn about the "circular financing" in the U.S. AI industry: funds circulate within the same financing ecosystem rather than coming from real commercial revenue. In contrast, leading Chinese developers mainly rely on internal cash flow.

Data richness is another advantage. Short videos are popular in China, providing local companies with a wealth of training materials. According to AI analysis network data, apart from two models from Google, all the top 15 image-to-video AI generation models globally are from Chinese companies.

Chinese companies also saw a rapid increase in the number of patents obtained in the U.S. last year, with Huawei ranking fifth among all commercial companies.

According to Fan Zhiyong, vice president and head of intellectual property at Huawei, the company's global patent licensing income reached a historical high of $630 million last year. He said that customers are increasingly favoring patents related to 5G and Wi-Fi, and show strong interest in the multimedia field.

All these efforts are driving an increasingly growing trend. (Translated by Wang Diqing)

Original: toutiao.com/article/7582798442544579126/

Statement: This article represents the views of the author himself.