This report offers a breakdown of the Sino-American AI competition by examining the two nations’ developmental ecosystems, the competition within them, and how industrial policy is playing a role. This is followed by an analysis of some of the emerging risks, particularly how AI will exacerbate modern dilemmas to threaten democracy and civil liberties. Finally, we consider some of the larger questions around AI, such as how to characterise its development, especially in a geopolitical sense, as well as the question of an AI pause.

• In “Structure of AI Industries” The report breaks down the structure of the ecosystems in which AI has developed, including their major actors, but also the geographical hubs, alliances, and academic institutes that shaped them.

• In “Impact of Industrial Policy” the report delves into the role played by Chinese industrial policy in driving AI progress, largely due to the government’s view that it is a core part of national development. Meanwhile, the US has given relatively small attention to industrial policy. Regulation in the US, meanwhile, is a growing topic of importance, despite its fractured nature.

• In “Deepdive: Sectoral Competition” the report highlights how the US and China are progressing in three specific fields of AI: Natural Language Processing, Computer Vision, and Autonomous Vehicles. While this a non-exhaustive exploration of competition dynamics between the two countries, it offers an insight into the followership and leadership taking place, and areas where each country has advantages.

• In “Horizon Scanning” we take a look at three risks that have emerged in the context of the Sino-American strategic clash. The risks created are by and large not new, and we highlight how AI will exacerbate and complexify risks already confronting the world today. These are a). Its impact on traditional arms racing, b) AI as a tool of mass surveillance, and c). AI as a tool for political disruption. In all three cases the team has outlined how AI development has increased the severity of already present risks. One of the out and out losers from this risk analysis is democracy and civil liberties, with Chinese firms ready to export AI to autocracies around the world.

• In “Considering AI Competition” the team examined two issues in the wider AI space, firstly, a conceptual one regarding the ‘race’ or competition itself. Often described as an arms race, we evaluate the argument that it should instead be termed an industrial revolution, finding it a much more useful framework to understand the complexity it brings. We then broach the issue of the AI pause and the need for regulation in this new technological space, advocating for this drastic, but potentially highly beneficial, move.