Is China prepared for the real impact of AI?
Amy Webb, the much-revered founder of the Future Today Institute, released her annual report on emerging tech trends for the year ahead on March 9.
China features heavily in the report, as it should, given its increasing impact and presence on the world stage, especially in technology.
However, as is typical with broad reports like these, many sections reveal a blinding bias. In the introduction, Webb makes a token reference to China as she tries to make sure readers know that the country will be in the report:
“It’s time to get comfortable with deep uncertainty. As I’m writing this annual letter to you from my office, we still do not know whether the UK will Brexit, if the special council investigation will incriminate President Trump or whether the 30th anniversary of the Tiananmen Square massacre will incite protest or apathy in China.”
While the first two examples could be innocuous statements of fact, the very structure of the third uncertainty reveals her bias, presuming that there will be—or should be—any reaction whatsoever.
Deeper into the report, in the section discussing artificial intelligence, she mixes the worst of China boosterism with shrill anti-China jingoism, claiming that China is on track “… to become the world’s unchallenged AI hegemon.”
The report even goes so far as to imply that China’s data surplus and investments in the sector are enough to allow the country to pull ahead of the rest of the world with “concentrated force and velocity.”
Love, hate
Opinions on China, rather like those of Donald Trump, tend to be extreme. Some love, others hate, few are in between.
In the case of tech, China is either turning into a “high-tech dystopia” where robots are taking over jobs and the government can track and punish you automatically.
Alternately, it’s about to win the tech Cold War because the country’s government and people are more “practical.”
The AI hegemon narrative is particularly popular.
In terms of AI readiness, China might be better off, but not by much. Sure, the country might lead in total investment into AI, and, while the country still ranks only seventh in the global AI talent pool, many estimates see them pulling ahead over the next decade. The government has certainly made many positive noises about the importance of AI, signalling support for the industry. The Made in China 2025 plan as well as a 2017 roadmap all place AI as a top priority.
A recent report by the China Institute for Science and Technology Policy at Tsinghua University, called the China AI Development Report 2018, highlights key trends in China’s AI industry. The report ranks China in the lead not only for AI research paper production (between 35,000 and 40,000 in 2017, compared to slightly less than 25,000 for the US) but also in the number of “highly cited” (the top 1% of citations received) and “hot papers” (papers published within the last two years that make up the top 0.1% of citations).
A 2017 McKinsey Global Institute report on AI in China shows that in 2015, Chinese research papers came in first for absolute number of citations, but quickly fell behind the US when accounting for self-citations (when a journal cites another article in the same journal).
The numbers are impressive—until you remember the many issues China has had with academic publishing fraud. For example, between 2012 and 2017, China retracted more scientific papers due to fake peer reviews than any other country.
We should be scared, but not because China is going to win the “AI race.” Rather, we should be very concerned about how unprepared governments the world over are regarding how AI will change society and the economy. While China’s legal system is still developing, Western democratic systems, especially in the US, have proven inadequate to deal with our current technological revolution.
In particular, politicians in the US continually show their ignorance. When Mark Zuckerberg danced around issues of privacy, his congressional interlocutors did not even recognize that he was doing so. More recently, Elizabeth Warren has called for the breakup of Google, Amazon, Facebook, and Apple. While there have been very compelling arguments as to why Google and Facebook should be considered monopolies, Warren’s targeting of Apple and their App Store shows how little she understands Apple’s true antitrust behavior(their rent-seeking via control of the platform).
While from a distance China’s progress in AI may seem enviable, the reality is that the country faces its fair share of challenges, too. Ambitious education projects and state-media that make sure the whole world knows the country’s AI goals will not actually solve China’s real problem of the forgotten heartland.
Missing in the heartland
As I have joked with friends on many occasions, Beijing and Shanghai are not China. The country’s coastal regions continue to reap the benefits of market reforms, innovation, and globalization with high GDP per capita. However, provinces such as Gansu, Yunnan, and Guizhou still lag far behind.
Even in more developed parts of the country, educational curricula still fall short of what many businesses need when it comes to talent. Some Chinese companies cannot wait for the education system to catch up, and are instead looking for ways to secure a pipeline of suitably qualified employees by designing and building their own courses.
SenseTime, in collaboration with the MOOC Center of East China Normal University, published China’s first-ever textbook on AI in 2018. The company promised that the textbook would be introduced to high schools around the country. However, as we’ve documented, the Chinese education system, both offline and online, is woefully underprepared to help train the next generation of workers, much less the current one. Cheap labor in China is already becoming more expensive, but very soon those jobs that in other countries helped push people into the middle class will no longer even exist.
The promise of artificial intelligence is already clear and visible: increased efficiency and productivity, fewer mistakes in critical areas of decision-making and diagnosis, and even perhaps greater efficacy in what the Chinese would call “social management.”
It appears neither China nor the US is ready for the real impact of AI—the coming dearth of meaningful work, and what that means for our societies.