Kai-Fu Lee is a Taiwanese-born American computer scientist, businessman, and writer based in Beijing. He has worked in China with Apple, SGI, Microsoft and, most recently, as President of Google China. He now oversees the Venture Capital fund Sinovation Ventures. In his recently published book AI Superpowers: China, Silicon Valley, and the New World Order, Lee describes how China is rapidly moving forward to become the global leader in AI, and is poised to surpass the United States largely because China’s vast demographics and system of top-down control are allowing it to amass huge data sets. Kai-Fu Lee has been dubbed, “the oracle of AI”.
On China’s rapid rise to AI dominance:
The US has better researchers, but “deep research” capability doesn’t necessarily translate to commercial value. China had new business models ideally suited to AI that were less developed in the US, UK and the EU. Think of shared bicycle services, and ordering relatively small food items or packages online. In these instances, AI combines with low delivery costs – for instance, around 50p per delivery cost.
On the ability of Chinese AI innovators to move overseas:
Tik-Tok, a social media video app for creating and sharing short lip-sync, comedy, and talent videos, is a good example, and doing well in South East Asia, Africa and now in the West too. Chinese companies are more willing to customise their AI apps and other systems to local needs. For instance, when Alibaba sold Alibaba Cloud in the Middle East they made changes and didn’t expect the customer to adapt, which is more common with US and European systems. Amazon Cloud Drive is far more reluctant to make these kind of changes.
On why China is not simply copy-catting or imitating in AI, but genuinely developing new methods:
China’s AI start-ups have phenomenal access to data sets. Tencent and Alibaba have some of the most powerful datasets anywhere. Consider mobile payments: These two companies alone have over a hundred times more data than PayPal does, and probably ten times more data than either Mastercard or Visa, while Mastercard or Visa don’t know how to use AI in the same way. China is “the Saudi Arabia of data”. Currently in China it is easy for both the state and corporations to collect vast amounts of data that is not so easily garnered in western countries.
On data privacy in China and its impact on AI development:
Lee believes that there will eventually be a General Data Protection Regulation (GDPR), similar to that in the EU, introduced in China. Most likely it will be weaker but will still work towards protecting consumers from bad behaviour by companies. When it comes to transferring data to third parties – a much more contentious issue with consumers – China already arguably has laws that are stricter than America. Selling data is a criminal offense in China.
On what UK AI entrepreneurs can learn from China:
Everyone has to accept that there are differing forms of entrepreneurialism outside the US and UK models. Specifically, there’s a need to learn to appreciate fast execution to get to market faster. Business optimisation is as important as creativity. We saw this in semiconductors where Taiwanese and Chinese chip makers are seen as “grunts” rather than hard workers.
On the prospects for European and UK AI development:
While the UK, Germany and France still have strong research track records, Europe has none of the success factors of the US or China. Europe has no venture capital entrepreneur ecosystem. UK and European entrepreneurs are not as innovative as America or as tenacious as China. Europe is strong on hardware and telecoms, but far less strong than either the US or China on developing consumer internet companies, social media companies, or huge mobile application companies. So far nobody in the UK or Europe has experience of working with large data or large scale AI companies such as you find in China.
On what UK and European companies can do to improve their AI entrepreneurship:
We can admire the idealism in the UK and Europe and the efforts to protect people’s privacy are admirable. But there is a need to build a brand new application and to focus on developing that. It can be done, it’s just a matter of using idealism to find the next paradigm shift and apply the innovation there.
On the future Sino-British relationship in the field of AI:
The UK clearly has world-class research institutions. Remember that the UK’s top AI company, Deepmind, beat the world champion at the ancient Chinese board game Go using its AlphaGo programme in 2016. This amazing achievement partly inspired many young researchers in China to go into AI development. If the UK is to keep its position as a leading research hub for AI it will need increasingly to build bridges with counterparts in China and combine China’s hyper-engaged users and vast oceans of data with the cutting-edge research happening in the UK.