US vs China in AI — A Realistic, No B.S. Assessment
Beneath the media hype this is the true picture of where the top two economies stand in AI.
With headlines like the ones below appearing in respectable publications, one could not be faulted for being misled into believing that China might be ahead of the US in artificial intelligence (AI) already.
If you think of China as a country that copies rather than innovates-think again. PHOTO CREDIT: Getty Images Living in…inc-asean.com
4 min read While everywhere else there is a fear of technology taking over people's job, China, the world leader in the…www.entrepreneur.com
Artificial intelligence is being used for a dizzying array of tasks, but one of the most successful is also one of the…www.theverge.com
But is it really? Let’s take a closer look beneath the headlines and the hype…
Quantity does not equal quality
(Yicai Global) Nov. 14 -- China has applied for more than 76,876 patents in the artificial intelligence sector to date…www.yicaiglobal.com
Yes. China has applied for a lot of AI patents and the rate is growing at the fastest in the world. But if you read this article closely, it also goes on to point out, “Though China has made more applications, it still has fewer high quality, high-value patents than world leader US and Japan…”
Companies looking to secure high-value intellectual property typically apply for an international patent via the Patent Cooperation Treaty, where the US makes up 41 percent of all applications — more than any other country — indicating its national strength in the sector.
— Chinese Firms Apply for More AI Patents Than Any Other Country, Yicai Global
And in terms of turning those patent applications into actual businesses generating significant revenue and brand power, China is still very far behind. According to research done by one of China’s ‘Ivy League’—Tsinghua University, only one Chinese entity ranks among the top 10 global AI patent owners — the state owned power grid enterprise.
Talent is still hugely lacking and lagging
China’s AI workforce is also still far behind the US in both quality and quantity.
China has 18,232 AI talents, compared to 28,536 in the US. More importantly, only 5.4% of the Chinese talents are considered ‘outstanding’, whereas the proportion is 18.1% in the US.
In case you think the study might be biased, again the numbers were counted by Tsinghua university.
To dig further, in a survey done by China Money Network, the founders of China’s top 50 AI companies were mostly educated abroad, including Harvard, Brown and several others in the US.
Funding vs Value created
Here’s where the picture starts to look a bit bleaker for the US.
Research firm CB Insights found in their 2018 AI Trend Report that China AI startups overtook the US in funding amount in 2017 for the first time ever, spiking up to 48% from just 11.3% in 2016.
To be fair, the US still has a lot more deals than China — 50% of the global total compared to China’s 9%. So that means the US still has a lot more AI startups being funded, though clearly China saw some very large deals in 2017. Otherwise, it wouldn’t have been able to account for nearly half the funding amount with just 9% of the global deal volume.
But that dominance in deal numbers is also trending down for the US, as shown by this chart below:
And here’s a less talked about fact that should make the Americans worry a bit. Almost half the money invested into AI by the three China tech giants — Baidu, Alibaba and Tencents —over the last four years are in US startups.
Nonetheless, even though China is clearly still some distance behind the US now in AI, it doesn’t mean it can’t catch up, or even overtake.
Indeed, if anything, money might be a key factor in China catching up.
“We often believe that the US is so far ahead, that the US invented everything…And it has dominated the tech scene, but a miracle happened 10 years ago. That miracle is money, of course. The Chinese government has vowed to be become a world leader in AI by 2030 with its strategic roadmap…Entrepreneurs and VCs started pouring cash into new AI startups.”
— Kai-Fu Lee, speaking at the O’Reilly Artificial Intelligence Conference in September 2018
Lee was formerly the president of Google China. He is now one of the most famous venture capitalists in China. Wired magazine calls him “something of a rock star in the Chinese tech scene, with more than 50 million followers on social networks within the country”.
Lee has helped launch five AI companies now worth a combined US$25 billion. He’s written about it in a new book call ‘My Journey into AI’. In his book, besides money, he said that China also has the advantage of very hungry entrepreneurs and abundant data on their side in the AI race.
His old boss at Google seems to agree with him. Eric Schmidt, CEO of Alphabet (Google’s parent company), said in November 2017 that the Chinese central government’s AI development strategy should “set alarm bells ringing in America.”
“… the U.S. needs to get our act together if it doesn’t want to fall behind on the technology that could determine the future of both the defense and commercial sectors… It’s pretty simple. By 2020, they will have caught up. By 2025, they will be better than us. By 2030, they will dominate the industries of AI.”
— Eric Schmidt
Maybe it comes as no surprise that the folks from Google are the ones ringing the alarm bells. It could be their own regrets knocking on their conscience. Their China equivalent, Baidu, started working on AI long before they did…
So the bottom line is this: China is still lagging in innovation and talent. But with money, highly-driven young entrepreneurs and government support, China has a good chance to dominate the world in AI.
Most of the top AI startups are focused on software aspects like natural language processing, computer vision and voice recognition. Given my own limited experience with some of their products, I would say their technology is as good if not better than western competitors already.
Here’s two more useful infographics on the valuations and spread of the top Chinese AI companies.