2024-04-27 07:18:05
Governments of each country and companies compete for AI talent
Mobilizing the financial power of American companies such as Meta-Open AI… Recruit talent by offering millions of dollars in annual salary
There are 2,000 AI-related departments in China alone… Produces approximately half of the world’s best human resources
Big Tech, war to secure talent in Europe and Japan… Non-AI human resources-field restructuring is not required
《US-China-Japan-Europe, AI talent recruitment war without gunshots
Major countries and large corporations around the world, including the United States, China, Japan, and Europe, are engaging in fierce competition to secure talent in the artificial intelligence (AI) field. To be selected as an AI talent, you must have an annual salary of several million dollars and a lot of trouble.》
“The battle for artificial intelligence (AI) talent is the craziest war we’ve ever seen.”
This is what Tesla CEO Elon Musk left on social media X (formerly Twitter) on the 4th (local time). He announced that he would significantly increase the salaries of AI technicians to prevent the turnover of AI talent within Tesla who are leaving for competitors. He pointed out that the main constraint slowing down the pace of AI development is ‘lack of talent.’
Developing and securing AI talent is not just Musk’s concern. As the world puts their lives on the line to secure AI leadership, the price of AI talent is skyrocketing. In the AI industry, there is a common saying that ‘those who gain talent gain models, and those who gain models gain the world.’
Due to this situation, not only companies but countries around the world are also competing to acquire top-level AI talent. The United States, which is competing for technological hegemony, has enormous financial power, and China is developing and retaining AI talent through reform of its education system. With Europe, Japan, the United Arab Emirates (UAE), and Singapore also participating, some say that the competition among countries to secure AI talent is reminiscent of a ‘war without gunfire.’
● The United States attracts talent with financial power
The secret to the United States leading the AI field with generative AI such as ChatGPT is overwhelming investment. According to the ‘AI Index 2024’ published by Stanford University’s ‘Human-Centered AI Institute (HAI)’ on the 15th, the United States ranked second with private investment in AI last year at about $67.2 billion (about KRW 92.67 trillion). It is significantly ahead of China (USD 7.76 billion). The cumulative investment amount in the United States from 2013 to last year is also more than three times that in China.
According to a recent survey conducted by Chinese science and technology media outlet ‘Taimaiti’ on the talent recruitment costs of 16 American and Chinese AI technology companies, American companies spent $770 million last year to secure core AI talent. wrote. On the other hand, Chinese companies’ spending was only $60 million. It is less than 1/10th.
If you look at individual companies, the differences become more stark. Meta, which is credited with sparking the recent ‘AI talent war’, pays its AI talent up to $2.51 million (approximately 3.46 billion won) per person, including basic annual salary, stock options, and performance bonuses. Open AI, the developer of ChatGPT, is courting talented individuals by guaranteeing $1.56 million, and Google is also guaranteeing $1.57 million.
On the other hand, Chinese information technology (IT) companies cannot provide such high salaries. ByteDance, TikTok’s parent company, pays up to $460,000, and Meituan, China’s largest delivery platform, only pays about $250,000.
It is not possible to hire talented people just by giving them a lot of money. “Key talent consider the computing infrastructure available to them to be an important factor in choosing a job,” Taimati said. Due to U.S. regulations on semiconductor exports to China, Chinese companies are currently unable to build equipment that satisfies AI talent.
In fact, in January of this year, Meta announced that it would purchase 350,000 units of NVIDIA’s high-performance AI chip H100 by the end of the year for AI research, which is interpreted as a ploy to recruit talent. Aravind Srinivas, CEO of Perplexity, one of the emerging companies in the generative AI industry, said in a recent interview, “We wanted to hire someone from Meta, but we were turned down because we don’t have as high-performance graphics processing units (GPUs) as Meta.” He complained.
● Aiming for a turnaround by growing talent pool
China announced its ‘Next Generation Artificial Intelligence (AI) Development Plan’ in July 2017 and set the goal of “becoming number one in the world in all fields, including AI-related academics and technology, by 2030.” Starting the following year, more than 2,000 AI-related departments were created at each university. Of these, more than 300 are operated by prestigious universities such as Tsinghua University and Peking University.
The effects are appearing quickly. According to the ‘Global AI Talent Trends’ published by Macropolo, a Chicago-based think tank, at the end of March this year, 47% of the world’s excellent AI talent (top 20% based on undergraduate degrees) in 2022 were produced in China. This means that one out of two AI talents attended an undergraduate degree at a Chinese university. The proportion of people from China increased by 18 percentage points compared to 2019 (29%), three years ago.
During the same period, the proportion of talent from the United States decreased from 20% to 18%. Even when narrowing the scope to top talent (top 2% of undergraduate students), the proportion of Chinese researchers increased by more than 2.5 times in 2022 (26%) compared to 2019 (10%).
The core of Chinese AI education is diversity and expandability. A representative example is the university’s ‘AI+X’ led by the Ministry of Education in 2018. This is a method in which students with basic knowledge such as physics and medicine integrate AI with their major fields. We considered that AI would be integrated with various industrial technologies.
Tsinghua University, a prestigious engineering school, created the ‘AI Class (Zhiban·智班)’ in 2019. It was named after the last letter of the name of Tsinghua University professor Yao Chi-chi (姚期智, 78), who won the Turing Award, known as the Nobel Prize in computer science. Only the best talent in each department, including mathematics, physics, and electronic engineering, is selected and AI-related training is provided separately. Even in the AI field, Professor Yao has a great determination to “raise geniuses like Albert Einstein.”
In addition, Peking University has the Department of Intelligent Medical Engineering, which combines medicine and AI, and Harbin University of Technology has the Department of Smart Vehicle Engineering, which specializes in autonomous driving.
There are many Chinese talents in major U.S. AI companies and research institutes. As of 2022, 38% of AI talent working in the United States were from China. This is actually more than those from the United States (37%). The New York Times (NYT) diagnosed, “The United States is unable to prevent the influx of excellent Chinese AI talent and is faced with the difficulty of having to prepare for China’s espionage activities.”
● Europe and Japan also join
Competition to secure AI talent is fierce even in Europe and Japan. In particular, as American AI companies expanded into various parts of Europe, everyone began recruiting talented people.
Open AI opened offices in London, England, and Dublin, Ireland last year. Microsoft (MS) also announced on the 8th that it would open ‘MS AI London’ in central London.
OpenAI has also decided to open its first Asian office in Tokyo, Japan this month. OpenAI CEO Sam Altman is known to be discussing with Japan’s Softbank to raise funds needed to build an AI semiconductor fab network. The Nippon Keizai Shimbun reported on the 22nd that Softbank will also invest an additional 150 billion yen (approximately 1.037 trillion won) by next year in infrastructure facilities necessary for the development of high-performance generative AI.
The rise of AI in each country is naturally leading to an increase in the price of talented people. According to Reuters, the basic salary of executives at AI companies in the UK has recently increased by an average of 50,000 to 100,000 pounds (approximately 85 to 170 million won).
Companies such as Meta and Open AI are not only offering salary packages worth millions of dollars to key talent, but are also introducing a method of evaluating mid- to long-term performance and rewarding them with stocks in the future to prevent job turnover. It was also widely known that Meta CEO Mark Zuckerberg and Google co-founder Sergey Brin contacted general researchers one by one to persuade them to change jobs or stay.
● Restructuring of non-AI sectors
Investment in AI core talent leads to workforce reduction in other fields. This is because big tech companies are willing to restructure their existing workforce in order to invest all of their resources into AI development and recruiting key talent.
Tesla, which is experiencing sluggish sales, announced to its employees in mid-month that it would cut 10% of its global workforce. Apple and Google’s parent company, Alphabet, also decided to lay off about 500 to 1,000 employees. Google CEO Sundar Pichai mentioned several businesses, including AI, as priorities to employees in January of this year and said, “The reality is that we have to make difficult choices (reduction in employment) to secure investment capabilities in these areas.”
Reporter Kim Cheol-joong tnf@donga.com
2024-04-27 07:18:05