Artificial intelligence, which is driving the fourth industrial revolution, is inseparable from geopolitics, and several world powers are vying for the lead. But what exactly are they competing for? Let’s take a closer look at the four key resources driving technological – and ultimately, geopolitical – advances.
Back in September 2017, Elon Musk, the founder of SpaceX and Tesla, tweeted: “China, Russia, soon all countries w strong computer science. Competition for AI superiority at national level most likely cause of WW3 imo”. World War Three may not have broken out (yet?), but neither does artificial intelligence (AI) form a purely scientific domain, detached from material concerns or geopolitics. On the contrary, it is an essential topic of national strategy and security.
Europe is lagging behind, its AI companies dwarfed by the GAFA giants
In the impact paper I wrote for ESCP Business School’s Geopolitics and global business impact series, I explain that the development of the AI industry is disparate across various nations. Unsurprisingly, the world’s major economic and political powerhouses are taking the lead, but they enjoy different competitive advantages. The US has cutting-edge hardware, research and talents, while China has a vast market and AI-ready data at its disposal. Clearly displaying its geopolitical ambitions, China has made the Digital Silk Road a key aspect of its Belt and Road Initiative and is investing significantly in the uncharted “Blue Ocean” of the African continent. Meanwhile, Europe is lagging behind, its AI companies dwarfed by the GAFA giants, each worth close to $1 trillion.
For countries aiming to become a global AI leader, the four following aspects are most strategic:
- Talent
So far, the US is the winner in the war for talent, with a pool of AI researchers numbering roughly 78,000 – twice as many as China.
Examining the dynamics of this resource, Jean-François Gagné’s Global AI Talent Report distinguishes “producer countries” with strong outflows (India, Israel) from “anchored countries” with stable talent pools (Japan, Russia), “platform countries” with high flows both ways (the UK, China and Canada) and “inviting countries” with strong inflows (the US and France).
How can governments attract and retain talents from abroad? Through liberal immigration policies targeting skilled workers, as Canada, Australia and the UK have set up, as well as with top-tier research laboratories and internationally-recognized educational institutions, as in the US.
- Data
Data may be the new fuel powering AI, but they are of direct use only after having been properly organised, processed, cleansed and analysed – which requires analysing techniques. For instance, ubiquitous online photos became valuable once computer vision and deep learning techniques matured.
Is Europe shooting itself in the foot with its strict data privacy regulation (GDPR)? In the global geopolitical battle, it risks falling behind the US and China, where the local tech giants (respectively Google, Amazon, Facebook, Apple, Microsoft, IBM and Baidu, Huawei, Alibaba, Tencent, Xiaomi) are able to massively collect and analyse data to offer their users targeted products and services.
- Microchips
Specialised AI chips are essential for deploying AI at scale. In that domain, the US and its allies dominate production by dominating the design of AI chips, though China is investing heavily and likely to catch up. Still, neither giant can maintain a full supply chain on its own and still relies on South Korea, Japan and Taiwan for contract manufacturing. The EU has seen its share of the chip market drop from 40% in the 1990’s to less than 10% nowadays, and is unlikely to regain an edge.
- Energy consumption
The ever-growing energy consumption required to run and improve AI is downright shocking. For example, training the MegatronLM language model consumed 27,648 kilowatt hours in nine days, almost three times the average annual consumption of a US household. Governments cannot afford to overlook the issue of energy consumption, which is a major factor of climate change, and must cooperate to develop a more sustainable AI ecosystem.