Internet Giants Caught in Tug of War for AI Talent
Wang Siqi
/SOURCE : Yicai
Internet Giants Caught in Tug of War for AI Talent

(Yicai Global) April 5 -- "Microsoft's artificial intelligence (AI) division currently has more than 7,000 engineers and scientists, and will recruit more, because Pony [Ma], Robin [Li] and Jack [Ma] are all poaching staff from us," joked Shen Xiangyang, also known as Harry Shum, senior vice president of Microsoft Global and head of Microsoft's artificial intelligence research unit, at a recent information technology summit.

As AI technology development entered into fast track phase, attracting talent has become the main competition area between the tech firms. Robin Li, co-founder of Chinese search engine Baidu, and Pony Ma, founding president of Tencent Inc., did not engage in a head-to-head argument during the IT summit, but they are actually in a tug-of-war for talent behind the scenes.

Competing for Talent

"[In the AI industry], for the first time in history, China now has a big market, as well as capital, companies and talent, and I firmly believe that AI will be main and most important battlefield for international competition in the future," predicted Zhu Min, an economist, at the summit.

Public data show that the US now has 2,905 AI companies, far outstripping China's AI pool of companies, which stands at 709 AI firms. However, China is on track to beat the world's largest economy in terms of the number of AI professionals.

Furthermore, 28.2 percent of the roughly 20,000 most important AI documents produced worldwide between 2005 and 2015 were written by Chinese authors, and their citation rate is as high as 32.1 percent. Google's much-discussed AI program, AlphaGo, was also first conceived by a Chinese scientist.

Chinese scientists laid a solid foundation for Chinese firms to compete on the artificial intelligence market, and Baidu, Alibaba and Tencent, collectively known as 'BAT,' have set up dedicated AI research units in the US to attract top talent.

Pony Ma did not deny the alleged 'staff poaching' Shen Xiangyang referred to, but said, "We also recruited many people last year and opened a lab in Seattle, the US, where Microsoft's headquarters are based. Many Microsoft employees are unwilling to leave Seattle, so we set up a lab in the city. That's how you can attract talent. You can't change it."

At present, more than 90 percent of the 50 staff members at Tencent's AI lab hold a PhD degree in artificial intelligence and have studied abroad.

"China wants to play a leading role in the AI industry. In terms of enterprise development, the largest AI companies, especially the pioneers of AI platform services, outperform Chinese enterprises in terms of strategy implementation and investment, but China has huge market opportunities emanating from its massive talent reserve," said Zou Shenglong, chief executive of Xunlei Ltd., during an interview with Yicai Global.

Over the years, Baidu recruited top-tier Chinese AI engineers and scientists and succeeded in laying a solid basis for in-depth AI technology development. It has a team of AI deep learning engineers previously headed by chief scientist, Andrew Ng. He was recently replaced by chief executive Lu Qi, but Baidu still remains at the forefront of deep learning technology in China.

With the launch of Tencent's AI Lab one year ago, the AI competition between the three leading Chinese internet companies has intensified.

'Fine Art,' an AI system developed by Tencent, won the 10th Computer Go UEC Cup after a stunning 11-game winning streak. Admittedly, Tencent is still lagging behind Baidu in deep learning technology, but it is aware of the enormous potential AI has and has started moving in that direction. As pointed out by Pony Ma, "Robin Li is leading the way in AI development. Tencent has joined in, but it's still lagging behind."

Pioneers of AI Commercialization

"Until 2013, we had been telling people that Baidu is a commercial undertaking, and shouldn't set up any research institute or any purely research-oriented organizations. But with the rise of deep learning technology, I felt that it was something totally new. It requires a lot of long-term strategic planning and theoretical and algorithmic breakthroughs. Starting from that moment, we began to invest heavily in AI talent recruitment and algorithm development," said Robin Li at the summit.

Microsoft, Google and trio of BAT are racing against each other to develop platform-level AI technologies and products. Small- and medium-sized enterprises have to wait for the major players to improve their technologies and platform. In the meanwhile, they need to penetrate the AI market as early as they can to get a competitive edge in product commercialization in the future.

"Artificial intelligence businesses are divided into two categories – AI platform builders and AI technology users. The first focuses on resource integration. Companies such as the BAT can mobilize the best AI scientists in the world to continuously develop new algorithms, and export their AI technology to other enterprises. The second, on the other hand, specializes in using AI to improve and even upset existing product models, including user experience, operational efficiency and structure," Zou stated. 

Companies in the first category are at a transitional stage as they still need to find the tools to make breakthroughs in commercial applications of AI technology, he continued, saying that it presents an ideal opportunity for small businesses. Operations at Xunlei's AI lab, on the other hand, focus on the development of commercial products. "The most important work for the lab is to identify the best 'formulas' available in the world, and turn them into tools. We apply them to existing scenarios and apps we develop, similar to the work of a chef," he opined.

The development of AI technology partly depends on the openness of AI platforms developed by the 'platform builders.'

"Artificial intelligence is a big business and will last for a very long time. We believe that it will maintain rapid development during the next 20 to 50 years. Obviously, no company can do everything alone to keep the momentum going. We open up our platforms so that other companies, especially those that lack necessary resources and long-term research capabilities [to develop the platforms], can focus on things that they are best at. This will benefit both sides," Robin Li underlined.

Follow Yicai Global on
Keywords: AI , PLATFORM , Baidu , TENCENT , Alibaba , DEEP LEARNING , Talent