This article explores three major challenges actors face when defining and carrying out their policies and answers in terms of high performance computing power (HPC) and artificial intelligence (AI), considering the political and geopolitical consequences of the feedback relationship linking AI in its Deep Learning component and computing power – hardware – or rather HPC. It builds on the first part, where we explained and detailed the connection between AI and HPC, and on the second part, where we looked at the related political and geopolitical impacts: what could happen to actors with insufficient HPC in an AI-world and the distribution of power resulting from AI and the threat to the Westphalian order.
Faced with the hurdles and threats stemming from inadequate HPC for the creation of AI-systems for AI-governance and AI-management, and, in a lesser way, for the training of these AI-systems, actors must devise responses. As they decide upon objectives and then ways to practically carry out responses, actors will face three supplementary challenges. First, objectives, planning and implementation regarding HPC must be thought in relative terms. Second, they must be seen envisioned dynamically. Third, the actors must consider that the very HPC field and thus the capabilities that need to be acquired are profoundly evolving because of the very feedback relationship between hardware and deep-learning we identified in the first part of our series “Artificial Intelligence, Computing Power and Geopolitics”.
Let us explain further each of these elements, while giving concrete examples for each. Using latest available data, the cases of Russia, with possible consequences for its intelligent android robot FEDOR, and Saudi Arabia will illustrate the importance of understanding relative HPC for AI. The importance of the dynamic element will lead us to a deep dive, including in terms of cost, into the race for HPC involving notably the U.S. and China, and to underline how this very race is a strong instrument of influence, wealth and power for those at the very top of the race, notably the U.S. and its companies, with China trying to catch up. Yet, as a result, the race also works hand in hand with the AI quest for optimisation to create an overall very fluid and revolutionary HPC environment.
We are thus faced with a series of feedback loops or rather spirals involving HPC and AI-systems and their developments, foreign policy, national interest and balance for power, defence, trade, ideology, business strategy and quest for profit, which, permanently, impact the field. The easy and apparently neat categorisation of the past are being erased. Similarly, possible responses, including one’s own, must increasingly be included within the foresight and warning, risk management or anticipation process when the main issue is AI and not separated from it. This is necessary to be able to properly consider how one’s strategy and action will impact reality and thus change the very range of future possibilities the initial foresight analysis considered. If we think about this two-fold evolution, there is nothing new, actually, but the speed at which events and dynamics unfold question the tidy distinction and especially the slow processes that were once presiding to polities and companies’ organisation. This is also one way AI fundamentally impacts AI-governance and Ai-management.
Now we have defined the complex framework within which actors must design their HPC policy, we shall look with the next article at the possible responses they may devise.
About the author: Dr Helene Lavoix, PhD Lond (International Relations), is the Director of The Red (Team) Analysis Society. She is specialised in strategic foresight and warning for national and international security issues.
Featured image: U.S. Army Acquisition Center – Nongkran Ch, Public Domain.