While US and China both aim to be world leaders in AI technology, they also both need to prepare for the impacts this investment in AI will have on their economies and workforces. Earlier this month, the Stanford AI Lab and Human-Centered Artificial Intelligence initiative hosted an exciting event focused on these topics as part of their recurring AI Salon series. The event involved a talk by Dr. Kai-Fu Lee1 outlining ideas from his book “AI Superpowers: China, Silicon Valley and the New World Order”, as well as a follow up discussion with professors Susan Athey2 and Erik Brynjolfsson3 concerning the ways AI will shape the future of work. The following video is a recording of the entire event:
The Strength of Chinese AI companies
Dr. Kai Fu Lee’s talk4 begins with US President Donald Trump saying “It’s a great thing to build a better world with Artificial Intelligence.” And then, the same Donald Trump-esque voice saying “AI is changing the world” – in Chinese! Both voice recordings were generated by AI speech synthesis technology developed by Chinese company iFLYTEK, and Lee highlights it as an example of how far AI technology as well as Chinese AI companies have come.
In his talk, Lee outlines his perspective on the “4 waves of AI”: internet AI, Business AI, Perception AI, and Autonomous AI. For each of these waves, he highlights a company that demonstrates the potential of that form of AI to revolutionize a particular industry. These include companies that provide micro-loans to individuals based on an AI-powered assessments, vision systems that can detect criminals as they enter crowded concerts, and finally autonomous cars and convenience stores. Furthermore, Lee argues China has a strong competitive position relative to the US when it comes to commercializing AI technology, for three reasons:
- There have been few truly significant AI breakthroughs: according to Lee, there has only been “one, single, big breakthrough” in AI – Deep Learning. Thus, even though the US and Canada have the most talented AI researchers, this does not translate to a significant advantage over China if Deep Learning continues to be the sole important innovation for commercializing AI; “Without big breakthroughs, it is hard for the U.S to retain its leadership because AI technologies are reasonably well understood,” Lee said.
- Openness of ideas and technology: Deep Learning has now been a popular topic of research for close to a decade, which has led to a plethora of open source code and free papers on the topic. Furthermore, the AI research community continues to publish new ideas on open platforms, and so no single country has an advantage with respect to availability of knowledge.
- Need for engineers rather than researchers: according to Lee, “We are now in the implementation phase. It’s a question of who can build the fastest … For most applications, you don’t really need super AI experts. Young AI engineers will suffice.” And, China has a quickly growing population of young AI engineers.
Chinese companies are further strengthened by their innovative capabilities, their tough entrepreneurs, their support from Chinese VCs and the government, and most of all – their access to data. According to Lee, “If data is the new oil, China is the new OPEC”, due to its huge population and heavy usage of data-generating service such as mobile payments.
AI and Job Displacement
Near the end of his talk, Lee notes that AI will also bring about many challenges5 in the video. In particular, it will displace many more jobs that it will create.
This is also the focus of the discussion with Susan Athey and Erik Brynjolfsson6. The first discussion topic questioned how broadly applicable modern AI algorithms are, and the panel’s consensus is that current technology is “narrow” rather than broad. Although researchers have recently solved challenging problems such as Go, these solutions cannot easily be generalized to tackling real-world commercial problems. Nevertheless, Lee estimates that approximately 50% of jobs 7 are likely to be automated with AI over the coming decades. Both Lee and Athey note government support for worker training is therefore important to deal with major job displacements, and Brynjolfsson highlights the importance of entrepreneurs creating new job categories that emphasize creativity over routine work. They further note that AI-powered automation may be an even greater challenge to countries that do not have as strong an entrepreneurial culture as the US and China.
The discussion then 8 turns to the question of incentivizing private companies to not only seek profit but also to create positive social impact. Lee states VCs have a role in finding and encouraging promising entrepreneurs that “have a big heart and see purpose beyond just making money”, and investing in pro bono work for social benefit. Brynjolfsson further notes that as a society, we all should recognize and celebrate people who improve society, rather than measure success solely by wealth. Athey then discusses the central role universities can play in this and provides the example of her own Initiative for Shared Prosperity and Innovation, which combines technological innovations for social good with design of market based incentives to enable philanthropists and non profits to subsidize new ideas and products that benefit society. According to Athey, AI is a great opportunity in this space because it has a lot of fixed costs but not a lot of marginal costs, so she is “optimistic about our ability to channel the philanthropists as well as the leading universities and research communities to try and tackle [social problems].”
AI and humanity
Both the talk and the discussion conclude with a focus on human life. At the end of his talk 9, Dr. Lee argues that AI will ultimately liberate people from routine work so they have more time to focus on living a meaningful life, and it is our responsibility to ensure that this will be the final outcome of AI development. Professor Brynjolfsson concurs 10 that AI is too often seen as the one making decisions about its own development, while in fact we as humans need to take active responsibility to direct how AI shapes our society. Lee ends the discussion with a touching note on what living a meaningful life means to him 11. While battling cancer, Lee realizes that he had spent most of his life optimizing for impact and success (much like an AI algorithm), but what truly makes him happy is spending time with his supportive and caring family. Lee contends that it is our capacity for love and meaningful relationships that truly make us human, and that it is up to us to channel the transformative forces of AI towards enabling us all to focus on our humanity rather than routine work.
Dr. Kai-Fu Lee is the Chairman and CEO of Sinovation Ventures and President of Sinovation Venture’s Artificial Intelligence Institute. Prior to founding Sinovation in 2009, Dr. Lee was the President of Google China. Previously, he held executive positions at Microsoft, SGI, and Apple. ↩
Susan Athey is The Economics of Technology Professor at the Stanford Graduate School of Business. Prior to joining Stanford, she was a professor at Harvard University. She is the first female winner of the John Bates Clark Medal. She currently serves as a long-term consultant to Microsoft as well as a consulting researcher to Microsoft Research. ↩
Erik Brynjolfsson is Director of the MIT Initiative on the Digital Economy, Professor at MIT Sloan School, and Research Associate at NBER. At MIT, he teaches courses on the Economics of Information and the Analytics Lab.He is the author or co-author of several books including NYTimes best-seller The Second Machine Age: Work, Progress and Prosperity in a Time of Brilliant Technologies (2014). ↩
6:30 in the video ↩
35:55 in the video ↩
47:00 in the video ↩
58:00 in the video ↩
1:15:00 in the video ↩
45:00 in video ↩
54:30 in video ↩
1:22:0 in video ↩