Huawei looks to challenge US rivals by promising big investment in AI talent

Why the Chinese telecom giant believes growing “human capital” is now more important than growing financial capital

"At Huawei we don't like the word transformation," the telecom giant's Rotating Chairman Eric Xu tells us via his interpreter. "What Huawei is doing isn't transformation because that implies you've changed from A to B. What we've done instead is made progress and moved forward."

Since its inception 56 years ago, artificial intelligence has succumbed to two winters where it suffered from a lack of funding and interest. However, as Xu rather poetically puts it, Winter always gives way to Spring, marking a new beginning for artificial intelligence and its pervasive capabilities.

The big announcements made from Huawei Connect in Shanghai prove that a lack of funding and investment will no longer be a worry for the AI industry. First was their full-stack, all-scenario AI portfolio which includes chips, chip enablement, a training and inference framework to "enable inclusive AI."

The ‘all-scenario' nature of the stack allows Huawei to facilitate different AI deployment set-ups; including public clouds, private clouds, edge computing in all forms, industrial IoT devices, and consumer devices.

On the back of this, the second big reveal came in the form of two new smart chips, Ascend 310 and 910, designed to natively serve all scenarios and allow for AI training at the edge or in smartphones, rather than in the cloud. However, despite a number of pre-emptive reports, Huawei are keen to stress they have no plans to corner the chip manufacturing market.

"As Huawei does not sell chips directly to third-party companies, if you're talking about pure chip-vendors, there is never competition between Huawei and them," Xu reassures the press. But it's a different if it's a hardware company or a cloud service company - "We provide hardware and cloud services so… there is competition here."

But the AI revelations don't stop at the full stack solution. There's also the announcement of the smart city development of Tianjin Binhai New Area in the North China Plain, the partnership with Audi to further boost autonomous driving efforts, an Atlas intelligent computing platform, the collaboration with Bosch to help boost the development of IoT in China… Well, you get the idea.

However, despite the shiny product and partnership launches, there is a larger, more serious overarching theme at this year's Huawei Connect. The current lack of artificial intelligence talent and what needs to be done to plug this ever-growing skills gap.

Over 22 countries across the globe have a national plan for artificial intelligence and $14 billion of venture capital funding went into AI-related investments last year. Despite this, figures also show that only 4% of enterprises have invested in or deployed artificial intelligence and of the global artificial intelligence talent needed, only 1% is currently available.

Xu used his keynote speech to highlight the investment gaps currently plaguing the industry: "To close these gaps, we need the right technology, the right talent, and the right industry ecosystem."

When it comes to AI talent, China is not doing too badly. The country attracted 60 per cent of global investments in artificial intelligence between 2013-2018 and accounts for 8.9 per cent of the world's total talent - around 18,200 workers.

However, according to research undertaken by Mai Mai, the Chinese equivalent of LinkedIn, talent in China faces the same regional distribution challenges as its Western counterparts. Beijing boasts close to 60% of all AI talent in the country, with Hangzhou, Shanghai and Shenzhen hovering around the 10% mark. No other Chinese city reaches over 2%.

Investing in future talent

While universities in China are working hard to plug the AI talent gap, education isn't always enough to ensure success. Hao Lu, the Chief Innovation Officer at Yitu Tech claims that most PhD students he interviews are actually less suited to work at the company than those who only hold a bachelor's degree.

"We choose not to hire those PhDs, not by decision but [because] they failed the interviews miserably," he explains. "Simply training a model is not something that difficult. With all the tools that are available, training a model is not super-hard science - even undergraduates can do it."

Instead, Lu has found that PhD holders are often missing attention to data, especially when it comes to analyzing it and applying understanding or intuition.

"Those are some of the things that are really missing when we interview a lot of PhDs."

This criticism from Lu is mirrored in comments made by co-founder of Mai Mai, Qian Wang, who believes the way talent is developed has to evolve in order to keep pace with the changing needs of the enterprise.

"Ten years ago, enterprises needed engineers, people who understood technology. Today they need scientists, because enterprise development now relies on capabilities like deep learning, cloud and IoT."

So, what's the solution? Senior Management Consultant at Huawei, Huang Weiwei's claim that 20-30 per cent of Huawei's annual R&D budget of US$15bn-20bn will be spent on research and innovation in an effort to attract more high-end talent is certainly a start.

It also goes without saying that if you want to attract high-quality AI talent, you need to offer good compensation and salaries. However, Weiwei acknowledges that money isn't always everything.

"These people don't come just because of the money - they want the company to have other attractive points. For example, the company should have enough investment and the right conditions to encourage R&D."

It's of little surprise then that the final announcement of the week came in a ‘AI Developer Enablement Program' shaped box. A promise from the telecoms giant to help train one million new artificial intelligence developers over the next three years by providing 20 hours of free, introductory online training, developer bootcamps and partnerships with established industry players.

A one billion-yuan ($143,936,700) fund has also been set up for universities and research institutes to help support AI talent development.

Zheng Yelai, who made the announcement on the last day of Huawei Connect, stressed that in the scheme of things, training one million AI talents is not an overly ambitious target; considering there are already more than seven million software developers in China.

The investments all clearly align with China's wider ambition of becoming a global AI powerhouse over the next decade; and the country's desire to challenge the US in terms of both technology and talent. 

"Managing cooperative relationships with our ecosystem partners is more important than managing Huawei's own business growth," said Huawei Chief Strategy Marketing Officer, William Xu.

"We believe that growing the industry and enlarging the market is far more important than increasing Huawei's own share of that market."