Why China May Be the Next AI Superpower, part II
The societal ripple effect
This is the second part for a recap and review of Kai-Fu Lee’s book AI Superpower. You can find the first part discussing the favorable Chinese AI climate after its Sputnik moment here (https://medium.com/value-stream-design/why-china-may-be-the-next-ai-superpower-d3451e0f95c6)
Why Chinese companies may yet conquer the world market
The unique characteristics of Chinese market and Chinese companies have made China a very difficult turf for many Western companies. Google and Uber are some of the known brands who have retreated from the market. Lee offers an explanation referencing different business tactics utilized by American and Chinese tech businesses.
According to Lee, most American companies are mission-based operations which look for a specific, grand change in the society and economy with their value proposition. They may sometimes hit incumbents hard, but the game is relatively fair and companies usually draw clear lines on what their business is and who their competition is.
When American tech companies enter a new market, such as China, they prefer to run the operations from US with minimal localization and integration to the local market. This way they can best achieve scale benefits and expand their revenue streams. The lack of interest to localize may be so deep that Lee blames many operations for being marketing-oriented (this is our brand, buy it) instead of what our brand in China should be.
The dynamics of the Chinese business development scene are described very differently. Lee even claims that hard working valley entrepreneurs are downright lazy compared to extremely hungry Chinese equivalents who work 12 hours a day, 6 or 7 days a week. How the Chinese companies operate in response is more important than the perks of their staff.
hard working valley entrepreneurs are downright lazy compared to extremely hungry Chinese equivalents
Lee repeatedly uses the metaphor of a gladiator arena to describe how startups struggle on the Chinese market. It is a bloody battle scene where hundreds of players, regardless of their relative weight, gather to battle for the rule of each market domain. Only a few remain standing in the end, and nobody leaves the field unscathed. The business moves executed along the way more resembling those of rivaling drug gangs than classy business men. The Chinese business landscape is the ultimate lean startup ground in that sense, requiring incredibly fast adaptation to market needs and pivoting the business to reflect the slightest change in customer preferences. The most fit will survive.
A key survival tactic exclusively practiced by successful Chinese companies is vertical integration, which most Western tech companies steer away from. By also binding into physical value chain, logistics, and custom hardware, Chinese companies can wall themselves from the competition. Everything in the digital domain is so much easier to copy that hardly anything there can save a business attacked by a better funded opponent that is willing to spend millions just to get rid of competition. This deep integration feeds back to support the data argument this story began with. By effectively connecting online and offline worlds with various IoT solutions, Chinese players accumulate the precious, big data about customers that will help them make AI solutions stick even better in the future.
The Fear of the AI World Order
Kai-Fu Lee is genuinely worried about the effects that upcoming AI-driven companies will have on society.
A full third of the book is dedicated to analysis of what may happen in the near future when US and China have pushed the current state of AI to new heights. Lee is not concerned with artificial super intelligence, or the singularity, instead he is very worried about unemployment and inequality which he projects will inevitably rise with AI. He equates AI as a new General Purpose Technology (similar to steam engine, electricity and ICT) and reviews the different projections about its expected impact on human employment, which range from 10 to 40% job displacement.
The prospects are grim anyway you look at it. Lee goes against some economists by arguing that this time, no hidden force is going to balance this development. He writes:
“free market is supposed to be self-correcting but these self-correcting mechanisms break down in an economy driven by artificial intelligence”.
The author sees that data-driven corporations have a tendency to create effective mono and duopolies against which traditional antitrust laws have little effect (although he acknowledges the European cases against Google). The accumulation of wealth and the decline of American middle-class are well-documented features of the ICT age which has otherwise stimulated GDP growth positively.
How will humanity prepare?
Because people in power of the global corporates are rational agents looking to maximize their returns, Lee doesn’t even consider the option of not utilizing AI to an increasing degree, which is the root cause of the situation. Because technological development is treated as inevitable, all there is to do is to treat the societal wounds afflicted by increasing equality brought about by AI.
Being solution-driven himself, Lee also briefly describes the current suggestions for fighting this development. These include reducing working hours, retraining workforce, and redistributing wealth. He does not buy into any of the existing proposals, such as universal income, but instead pitches his own idea.
In a quite radical turn justified by his personal experiences with cancer, Lee goes from economical growth to well-being and suggests prosocial ventures to balance the impact of AI companies that amass huge returns. These social ventures should aim for linear, not exponential returns and compensate everyone in the Missing text
Lee’s own thinking culminates in the following quote: “the synthesis which I believe we must build our shared future: on AI’s capability to think, but coupled with human beings’ ability to love”. He thus raises compassion as the key differentiator for
The obsession about social relationship goes back to his personal revelation about the power of love as well as the argument that for now, social interaction (as well as manual dexterity) is the furthest away from the scope of AI applications. It is thus likely to survive human-dominated for the foreseeable future (before we all dive into Matrix type of simulated life, a dystopian version of Ray Kurzweil’s future visions).
Social investment stipend is the core idea that Lee proposes to substitute AI-replaced occupations with. The proposal requests people to work for common good in duties requiring social interaction in exchange for monetary compensation.
Conclusions
I appreciate Lee’s thinking, but I fear it has many issues similar to universal income path he strongly criticizes. He acknowledges several problems with it himself, but I say in his defense that the most important part is that he as a respected thought leader among the world’s technologists has come forth by articulating problems and proposing alternatives to the prominent values in techno-optimist world view that most technology entrepreneurs show.
Lee can’t provide definite solutions, but he makes a convincing argument about the problem
As such, the book combines great insights as a primer to Chinese tech entrepreneurship, basics of AI-enabled business as well a short autobiography of the author.
How does it read?
The book is not an excessively long one, but still feels like three, not one book. The first part describing AI and Chinese entrepreneurial world contains much repetition and fewer insights. The plot thickens in the second part that contains many insightful case descriptions of how things in China lay and how its tech culture has evolved.
The third part is then almost like a new book. It goes on much like an autobiography in a stark contrast to earlier part which resembled a (colorful) McKinsey business report. Although the Lee point-of-view is maintained, the narrative jumps around.
It is worth noting that although book is namely authored by Lee, it appears to be for a better part composed by Matt Sheehan, a professional writer. This is openly acknowledged. Altogether this is not a literary masterpiece, but a feels like a quickly produced book about an extremely timely topic. The speed is likely the best explanation why it is slightly incoherent, tends to repeat itself and has untypical structure for a business book.
Short sighted
The greatest shortcoming of the book that attempts to both identify and specify a problem and its potential solutions is that it remains awfully short sighted. While I don’t want to downplay the seriousness of the issues Lee reveals, I fear that he is ignoring a big part of the world as well is its less mundane issues. There are pandemic issues that still plague both developing and industrialized countries (such as obesity, diabetes, malaria, cyber espionage, etc.) which could be resolved to a large degree if global capital and policy aligned any better. While global warming is mentioned as one of the things artificial super intelligence could solve, it would have served as a another real global threat that the global superpowers seem to be unable to handle as it seemingly is on the way economic development.
Looking at the world from Northern Europe, from a country not even mentioned in the book, I get a very depressing image out of the situation. As the book basically sees the whole global market as a target for US and Chinese companies to fight over, it seems rest of the world would only be served scraps. From the $15.7 trillion GDP growth predicted by PwC , could this really fall mostly into hands of two players? Particularly the mechanisms that are suggested to remedy the increasing inequality created by new AI industries seem to fit only the needs of these two aforementioned countries. How is the rest of the world supposed level this impact? Will this development actually turn Europe, most of Asia, Russia, and notable G7 countries back into developing nations? Hopefully not.
However, there’s nothing in the way of the development at the moment, as described by Lee. Instead, global investment money is eager to seed new AI-based technology success stories anywhere in the world. Chinese companies are spreading around the world with different strategies than American ones. In a few years we’ll know if Chinese and US ambitions in AI will be met with another AI winter, or if we’re looking at the dominating AI superpowers like those projected by Kai-Fu Lee today.
Finally, the author seems to be very discreet about discussion any notions of Chinese governmental oppression, surveillance and limitations of free speech (i.e. censorship) which can be more effectively enforced with new technologies. It is known that these policy decisions were partially responsible for Google’s first China retreat. As Google is once more trying to get things moving on the Chinese front, many of its employees have publicly rebelled against any co-operation with the government (e.g. search engine project received much bad press just in August https://www.bbc.com/news/business-45216554 & https://nordic.businessinsider.com/google-china-exec-human-rights-2018-8/ )
Learn more:
Kai-Fu Lees TED Talk from August 2018
https://www.youtube.com/watch?v=ajGgd9Ld-Wc
Future of Humanity Institute, Oxford University: Deciphering China’s AI Dream
https://www.fhi.ox.ac.uk/deciphering-chinas-ai-dream/