Global ambition behind Beijing’s bid to boost AI talent

China has initiated a multi-pronged effort to swiftly boost the training of artificial intelligence (AI) talent and AI research within its universities in order to prop up its ambitious plans for ‘global dominance’ in AI, backed by a huge injection of central government funding.

Beijing is also exhorting the country’s giant technology companies to co-fund the expansion of training in universities, including funding professorial chairs and training courses in universities, in order to accelerate the growth.

Joint projects between the government and the country’s high-tech industries will include a major scholarship programme for Chinese students to study AI at top institutions in North America and elsewhere overseas, particularly at the doctoral and post-doctoral levels, and a boost to China-United States educational exchanges in AI, as well as other types of international cooperation such as joint laboratories.

Last month China’s ministry of education released an action plan for higher education, to bolster AI innovation, calling on the country’s universities to become the “core force” for China to become a global AI innovation centre by 2030, as outlined in the New Generation Artificial Intelligence Development Plan, released by China’s State Council in July 2017.

The plan clearly lays out the internationalisation of China’s AI efforts, calling for Chinese scholars to “enhance their international influence” by taking up “important positions in international academic organisations”, while “actively participating in the formulation of international rules related to AI”.

However, its main thrust is setting targets to dominate AI globally by 2030. In the next two years “at least 50 artificial intelligence academic and research institutes” will be set up, according to the ministry. Already in the past year the Chinese Academy of Sciences, Beihang University and Nanjing University, among others, have established new AI degree programmes and institutes.

“By 2025, colleges and universities will significantly improve the quality of training in the field of new generation of artificial intelligence technology and innovation and will achieve a number of original results of international significance and some theoretical research, innovation and application of technology to reach a world-leading level,” the often bombastic ministry of education document states.

“By 2030, colleges and universities will become the core force for the construction of the world's major artificial intelligence innovation centres and peak talent to lead the development of a new generation of artificial intelligence.”

Overall funding has been provided to train 500 teachers and 5,000 students in AI at top universities over the next five years, the official China Daily newspaper reported.

Faculty bottlenecks

Chinese official media has hyped this as an attempt to build “the world’s largest AI talent training programme”, but significant hurdles will have to be overcome to achieve this, experts say.

“We’ve gone from a situation in a very short time where nobody thought of China as an AI player to a situation where there is this duopoly of China and the US within the space of the year,” notes Paul Triolo, an expert in China innovation at Eurasia Group, a political risk consultancy, who notes that there was little hype in China prior to the July 2017 AI plan’s release.

A recent report by Goldman Sachs, frequently quoted in official Chinese media, noted that although China accounts for “over half of the new AI projects in the world, it has merely 5% of the global AI pool”.

Despite significant rises in China’s AI research publications and patents, pushing it to the number two spot globally behind the US, the US has a decades-long head start within its universities and high-tech companies in Silicon Valley, with an established research base in the very fields China is hoping to dominate globally in just over a decade.

“The reality is that there are bottlenecks,” notes Triolo. “China is way down in having really good, robust university programmes that have sufficient trained faculty in that area.”

According to a report released in March co-authored by Triolo with Jimmy Goodrich, vice-president for global policy at the Semiconductor Industry Association in the US, less than a dozen universities in China have faculty trained in the field, compared to about 100 universities in the US.

The short-term outlook for China’s universities to deliver AI talent in significant numbers “is pessimistic”, Triolo recently told University World News. “Over the long term some of the structural support under the government’s AI plan will start to kick in, but it takes time to train people.”

‘Trade war’ threat fuels domestic training drive

To kickstart its AI drive, luring back Chinese students from overseas universities is a major focus for the Chinese government, while China’s big tech companies have been big buyers of technology companies in the US and elsewhere, in part to access research and development, patents and talent.

But a looming US-China trade war, sparked by President Donald Trump’s focus on the US trade deficit with China, has put Chinese companies, including tech companies, in US sights. This could affect Chinese acquisitions of US tech companies and with it access to top-end talent, as well as hit major collaborations of US and Chinese AI researchers.

“In the US, AI is seen as a strategic issue tied to national security, which it hadn’t been until more recently, which may have a dampening effect on some of this collaboration,” Triolo notes.

“The US-China tech cold war is heating up and will get worse, particularly in the next year, and one of the ways the US could respond could be to limit work visas for Chinese. It will be a much more hostile environment for some of these exchanges,” Triolo said, referring to reports that the US may restrict visas to Chinese researchers in sensitive areas.

Homegrown talent

With these threats, China has turned its attention more urgently to building up homegrown research talent.

The ministry’s detailed action plan calls on Chinese universities to develop interdisciplinary links between AI and subjects like computer science, mathematics, physics, psychology and sociology – the aim is to swiftly bring AI research into real-life applied technologies.

Ministry officials note that although at least 10 universities have set up AI-related majors, these are mainly theoretical and students lack “practical operational skills” and often cannot meet employers’ requirements.

The aim is to swiftly create job-ready graduates for the AI industry. The ministry of education on 3 April launched a five-year ‘International AI Training Program for Chinese Universities’ at Peking University (PKU) in Beijing, setting up AI as a university discipline, and in collaboration with tech companies and Chinese technology investment firm Sinovation will train “the first 100 teachers and 300 students” in AI.

Some 106 teachers from 49 Chinese universities were participating in the “first round of training” at PKU, the ministry said.

“The training programme will hopefully integrate education and industry resources from home and abroad to develop a universally-recognised education system with Chinese characteristics,” said Tian Gang, PKU vice-principal, at the official unveiling of the new training programme in early April.

PKU recruited faculty from abroad, including IBM Professor of Engineering and Applied Mathematics at Cornell University John Edward Hopcroft who has taught AI courses in China for a decade.

Xu Tao, director of the International Cooperation and Exchanges Department at the ministry, said AI would be “upgraded to a first-level subject” as a full academic discipline rather than being confined to a few courses as part of computer science disciplines.

Collaboration with industry

The obvious haste to build up a world-leading sector has also meant the government is putting pressure on its largest tech companies to invest in driving the pace to develop talent.

A great deal of AI research and training actually occurs within China’s big tech companies, particularly online retail giant Alibaba, social media and messaging app behemoth Tencent and internet search company Baidu – the three are among the world’s largest internet companies.

“These companies maintain extensive R&D facilities not matched within leading government research organisations such as the Chinese Academy of Sciences,” notes Triolo.

In the wake of government exhortations, Baidu CEO Robin Li, his wife Melissa Ma and the company jointly donated CNY660 million (US$104 million) to Peking University in an announcement last month. According to the statement, the new Peking University Baidu Fund will support cutting-edge research in applications of AI intelligence in the medical and social areas.

Liu Qiangdong, CEO of Beijing-based JD.com, another huge online retailer, and his wife Zhang Zetian donated CNY20 million to Tsinghua University, Beijing – recognised as having the country’s strongest AI research capability – for research projects in AI and logistics.

In March China’s Alibaba announced an initial five-year funding package for a new Alibaba-NTU Singapore Joint Research Institute at Singapore’s Nanyang Technological University “for the creation and testing of artificial intelligence solutions to address societal challenges”, with a pool of 50 researchers from Alibaba and NTU.

According to the announcement, they will also work to build a crowdsourcing platform “to connect researchers and industry practitioners around the world within an AI-focused R&D ecosystem”, and are encouraging global AI experts, research institutions and universities to join and contribute to the AI research community.