Wise Leadership and AI - 5 | The AI Race
The trend towards ‘techno-nationalism’, where a nation relies on its own resources to become a leader in AI machine learning and other crucial strategic tech areas, is fully underway. The race is on, but how is it playing out? And why will wise leadership talent be so important?
The groundbreaking deep learning approach to AI has turbo-charged the cognitive capabilities of machines. This ‘narrow artificial intelligence’ (versus the as-yet-unachieved ‘general artificial intelligence’) can now do a better job than humans in a range of domains, from identifying faces and recognizing speech, to issuing loans.
Unsurprisingly, many companies, and the countries from which they originate, are eager to master them. How these countries, especially the USA and China, the two most advanced in AI, choose to compete and cooperate, will have a dramatic effect on global economics, and geopolitics in general.
1 - The stakes for global AI pre-eminence are enormous
‘Winner-take-all’ and ‘first-mover’ dynamics in internet-based industries have huge potential to fuel the economic growth of countries. AI technology has a powerful role to play in maintaining social stability, from health and transportation, to the justice system and public security. For nations that want greater independence from the US engine, a strong AI base will help them decouple.
2 - China’s AI industry is under the spotlight
China aims to be one of the most advanced nations in AI by early 2021, leading in most fields by 2025, with direct AI output exceeding USD 60 billion. Some predict it will be the world’s primary innovation center by 2030. AI is a paving stone in the New Silk Road proposed by President Xi Jinping as part of the Belt and Road Initiative (BRI); trade networks connecting Asia, Africa, and Europe. The vision is of a world-class technological superpower, an international community of practice built around top Chinese AI talent, with China as a primary global AI exporter. This geopolitical re-positioning would allow China to cement partnership and export opportunities, advancing its technical and geopolitical leadership objectives.
3 - China has numbers on its side
Much AI development and discovery work was driven by elite researchers clustered mainly in the USA, Canada, Israel, UK and France. Now the age of AI implementation is dawning. Computing power and good AI algorithm engineers are vital. But beyond a certain threshold, data is key for power and accuracy. Regarding 4 critical success factors for AI implementation; data, entrepreneurs, AI scientists and an AI-friendly environment, China is well placed to outpace the USA. Its AI entrepreneurs are implementing data in multiple commercially-viable AI-powered applications, drawing on a domestic population of 731 million internet and 685 million mobile users in 2019, over twice the USA equivalents. China also has about 286 million digital natives versus 75 million in the USA. Thus, the struggle for AI supremacy will become fierce.
4 - Rising giants
US firms that widely use AI algorithms include Google, Amazon, Microsoft, Facebook and WhatsApp, but also Apple, IBM, Netflix, Airbnb and Uber or Lyft. Chinese-backed equivalents include Alibaba and its online shopping website Taobao, the Alipay payment platform (54% of the Chinese market), and Tencent, with its social messaging app WeChat and WeChat Pay (32%). Further examples include QQ Wallet, Baidu, or Didi Chuxing, the player that drove Uber out of China. In robotics, European firms — especially German - may still play a globally competitive role. Germany’s AI strategy plays to its industrial strengths, drawing on data in cyber-physical manufacturing and the industrial IoT.
5 - The Chinese government has a clear strategy
AI’s impact on the Chinese economy is estimated at US$ 7 trillion by 2030-2035 (a quarter of future GDP), contributing 1-2% to GDP growth p.a. The government has supported its “national champions” with substantial funding, encouraged domestic companies to acquire chip technology through overseas deals, and made long-term bets on supercomputing facilities. Baidu and start-ups like Cambricorn are designing chips specifically for use by AI algorithms. Its strategy embraces strong state support and intervention, the transfer of technology and talent, and investment over a long-term period.
6 - The battle between China and the US looks set to intensify
Geopolitics will be increasingly determined by the power struggles in the AI field. Yet, even if China has a numbers advantage over the USA when it comes to data as the new oil, the US still leads in the theoretical frameworks for deep learning innovation. And although China has the edge in face recognition algorithms, Google and other Silicon Valley players are still generally ahead of the AI game. Moreover, China still lacks the international data necessary to reduce biases for apps that could be used beyond its national borders. But there is everything to play for — and with.
7 - China’s AI talent is in demand
In 2018, China had 50,000 experienced AI professionals, versus 850,000 in the US. But China has a significant manpower advantage in training and labelling the data needed by AI applications. And it has been highly creative, with rapid user adoption and acceptance creating a strong feedback loop. Its fundamental AI research targets image and speech processing. In image recognition, China is now world-class, with leading companies in NLP such as Flytek, Baidu, Megvil (Face+++) and SenseTime.
Recent research highlighted a ‘global brain drain, US brain gain.’ Most upper-tier research happens at US institutions who rely on foreign-born talent for more than half of their output. Chinese researchers are the largest source of global AI talent. Most end up studying in the US, and working for US companies. However, there is evidence that ethnic Chinese researchers are returning from the US, and that China is turning the tables on the brain (and corporate) acquisition flow.
8 - China’s AI strategy has 4 factors – and it is gaining ground
- Hardware in the form of chips for training and executing AI algorithms.
- Data as an input for AI algorithms.
- Research and algorithm development.
- A commercial AI ecosystem benefiting from guidance funds set up by local governments and state-owned companies.
The Artificial Intelligence Potential Index (AIPI) estimates a country’s overall AI capability. A 2018 assessment put China’s AIPI score at 17/100, versus the US at 33/100. However, this is relative and individual companies may gain a specific competitive advantage, not quantified at the aggregate level. Just a year later, another report scored the US 44 out of 100 available points, China 32 and the EU 23. If the US (still) leads in theoretical potential, China is implementing at speed, hindered by fewer legacy systems.
9 - Is AI poised to bring about a New World Order?
Georgetown Professor Nicholas Wright distinguishes three facets:
- Political regimes: Liberal democracies could compete with authoritarianism.
- Social sectors: Practical AI applications will transform multiple, fundamental fields: from transport, to healthcare or defense.
- Singularity and the sense of self: AI ‘singularity’ will change our self-concept as humans.
AI applications provide a nation with a decisive strategic advantage in international security. Yet countries must prepare for answers to foreign cyber-attacks, beyond the “do-nothing or do-little-approach”. Several scientists see AI as part of a surveillance system, e.g. via Silicon Valley companies predicting and influencing behavior, while the government closely monitors its citizens. AI may become the newest tool to aim for some kind of global supremacy.
10 - Wise digital leadership is more important than ever
AI raises the possibility of social disorder and political collapse due to widespread unemployment and inequality between the AI ‘haves’ and ‘have-nots’. Ethical concerns require a robust civil society and in the growing battle between economic powers, a major risk is that a given party chooses to magnify the capabilities of AI over safety. Regulations and global governance may be needed (without stifling AI research), making social and political choices that align our interests as a community. Outsmarting the ‘Other’ will not be enough – leaders must be wise enough to prevent destructive forces.
We should understand AI’s limits and darker side and the scale, scope and speed of potential disruption. Only wise, measured and mindful leadership will be able to conceive of some form of global good for these innovative but ethically-neutral techniques, realistically implementing them in our specific world context.
Jimmy Chen, Managing Partner of Amrop China, suggests that “The New Intelligent Economy requires multi-faceted leadership talent; professionals with sound knowledge of technology, people and operations management. These people are in scarce supply. Even the most senior digital leaders are often only in their 30s. If senior IT talent is plentiful, then digital experts are a new breed, emerging only 5-6 years ago in China.”
The stakes for global economic pre-eminence in AI are enormous. ‘Winner-take-all’ and ‘first-mover’ dynamics in internet-based industries will enable successful countries to sustain economic growth, meet GDP targets, and become more independent of the US economy — assuming that such a decoupling would be possible or desirable.
Authors: Dr. Peter Verhezen, with Jimmy Chen, Amrop China.