China Aims for Global AI Leadership
Manny Frishberg and Mary Anne M. Gobble
Research-Technology Management. 61.3 (May-June 2018): p3+.
DOI: http://dx.doi.org.db06.linccweb.org/10.1080/08956308.2018.1445383
Copyright: COPYRIGHT 2018 Industrial Research Institute Inc.
Full Text:
Artificial intelligence, or AI–computers that can analyze data, make predictions, and draw conclusions in ways only humans could a few years ago–may be the most important technological development since the steam engine, movable type, or cuneiform writing on clay tablets, depending on who you listen to. “I have a hard time thinking of an industry we cannot transform with AI,” Andrew Ng, cofounder of Coursera and Google Brain, has said. Ng is now chief science officer for Baidu, China’s leading search engine company; he directs Baidu’s AI research effort in Silicon Valley. Ng’s move, from leading US tech companies to a Chinese tech powerhouse just moving into AI, is indicative of a wider shift. China, increasingly, is seeking to challenge Western dominance in AI.[unique_solution]
“Artificial intelligence” refers both to systems that show aspects of human-like intelligence–understanding language, reasoning, problem solving, and the like–and to the field concerned with the development of such systems. Recent developments in AI have largely been driven by advances in an approach called machine learning, in which a set of algorithms and devices enable the computer to parse large quantities of related data and draw conclusions by passing the data through layers of artificial neural nets–circuitry that mimics animals’ brain connections–asking yes/no questions to uncover patterns. Adding a feedback loop that assesses whether decisions emerging from this process are right or wrong enables “learning”–the system can use the feedback data to modify its approach to later iterations. Add enough layers of neural nets and feedback loops, and “deep learning” emerges, a version of unsupervised machine learning so complex that even the engineers who build and program these thinking machines cannot say exactly how the information is processed in those layers.
The United States has led the field since its inception, with US-based researchers and corporations making cutting-edge advances; most other countries have lagged far behind. China, however, is rapidly closing the gap, increasing both public and private investment in the field and generating a growing number of research papers. Until 2013, the United States led in the number of research papers mentioning deep learning and the number of citations in other papers, a measure of their importance. In 2014 and 201 5, China led the US in both metrics. In 2017, nearly as many papers were accepted for the annual meeting of the Association of the Advancement of Artificial Intelligence from Chinese researchers as from US-based researchers. The conference had to be rescheduled to avoid the Lunar New Year because too many Chinese researchers could not attend at that time.
These developments are part of a larger plan, which the Chinese government made public last July in its Next Generation Artificial Intelligence Development Plan. That publication outlined a three-stage plan: to be on a par with the most advanced countries in research and applications of AI by 2020, to be making major theoretical breakthroughs in AI and to a world leader in applications by 2025, and, according to the plan, “By 2030,
China’s AI theories, technology and applications should achieve world-leading levels, making China the world’s primary AI innovation source.”
The plan calls for significant government investment in theoretical and applied AI research while recognizing the lead role private companies have taken–and will continue to take–in creating and deploying commercial applications. The government is moving quickly to enact the plan. In January, Reuters reported that the Chinese government will spend 13.8 billion yuan ($2.12 billion) to build a “national-level” AI center in Beijing to house up to 400 AI-based enterprises. Chinese companies such as Didi (China’s closest analog to Uber) and Tencent, which makes the WeChat messaging app, have already opened their own AI research labs.
China’s pattern of using state-guided investment to support industrial policy is contributing to the rapid development of the field there. Access to the vast amounts of data deep learning requires is another edge the country has, a result not only of its industrial policy but also of its willingness to collect and use information about its people. Privacy issues that can hamper AI research and applications in Western countries are simply not an issue in China. The availability of large troves of data that can be used to train AIs and generate deep learning means that, Chinese companies are naturally positioned to lead in a variety of AI-related fields, including facial recognition, voice recognition, and natural language processing. US companies are also recognizing the potential–Google recently announced plans to build its own AI lab in China.
For the time being, the United States retains the edge in AI R&D. One important venue for the emergence of new technologies is the startup world. Although China’s economy is increasingly entrepreneurial, it still can’t match the US in this regard: Last year, there were 39 AI-focused startups in the United States, in more than a dozen fields; in China, there were just 3, focused on robotics, health care, and conversational AI and chatbots. And other US companies are joining the fray, most notably around autonomous cars, where a growing number of competitors across industries have entered the field, from Google and Tesla to Ford and Freightliner.
However, the United States is lagging significantly in government support for AI. The National Science Foundation’s 2017 budget proposal for Information and Intelligent Systems research came to just over $207 million, $12 million more than in the previous two years. But only a third of that, just 6.6 percent of the total NSF budget, is earmarked for “Robust Intelligence”–basic AI research. The Pentagon’s 2017 budget plans included a request for $12-15 billion to fund AI weapon technology.
It’s unclear how much of that funding will remain in the 2018 budget, especially for nonmilitary research. President Trump’s 2018 budget proposal slashes 11 percent from the NSF budget, which could amount to $22 million less for AI projects. At the World Economic Forum in Davos, Switzerland, in January, according to Zanny Minton Beddoe, editor-in-chief at The Economist, there was wide concern that the Trump administration is doubling down on manufacturing and natural resources while China makes both basic and practical AI research a priority. As a result, the United States and the West as a whole could lose its edge in a field that’s likely to shape everyone’s future.
Artificial intelligence holds the promise to reshape the world over the next century, much as computers and the Internet defined the last one. The Internet revolution was led by an American synergy between entrepreneurial ventures like Apple and Microsoft and the government research agencies whose basic work created the network’s precursors. Who will lead the AI revolution still remains to be seen–but China is determined to be in the running.
Manny Frishberg, Contributing Editor Federal Way, Washington mannyf@q.com
DOI: 10.1080/08956308.2018.1445383
Source Citation (MLA 8th Edition)
Frishberg, Manny, and Mary Anne M. Gobble. “China Aims for Global AI Leadership.” Research-Technology Management, May-June 2018, p. 3+. Academic OneFile, http://link.galegroup.com/apps/doc/A540678616/AONE?u=lincclin_dbcc&sid=AONE&xid=169c886d. Accessed 19 June 2018.
Gale Document Number: GALE|A540678616