There is No Such Thing as an AI Arms Race

This content has been archived. It may no longer be relevant

Although the competition for artificial intelligence leadership is not an arms race in the classic sense, the narrative of an „AI Arms Race“ has been popular, especially among US defence circles. Other people have thoroughly debunked this notion already, so I will not belabor this any further, except to point out a German-language article by Carlo Diehl and myself from last year. In it, we gave a brief overview and summary of US, Chinese, and EU policies for AI.

Carlo also came up with this masterful timeline of key documents and policies from 2016 to early 2021. Given how quickly the field develops, I wanted to share it here before it becomes completely obsolete.

I like this graph because it shows the density and rapidity of political action. It does not show (but suggest) how intertwined these national political processes were. Much of what we see here happened in reaction to one another, which makes this not an arms race, but an innovation race (as I argue in another – as yet unpublished – paper with Stefka Schmid and Christian Reuter). However, it’s not a „regular“ innovation race as described by innovation economics but a particularly geopoliticized one. States pursue technological leadership not primarily (or at least not only) for economic reasons, but also for the purposes of security and international status. This is what Kai Oppermann, Jakob Landwehr-Matlé and myself are trying to parse in our „Tech War“ narratives project.