Transformernews.ai reports that US technology companies have advanced the AI race with China narrative since 2017 to shape export controls, military partnerships, and lighter regulation.

SHShakeel@ShakeelHashim

It was so exciting to work on this paper with Seán, Yi-Ling and Coleman! Very fun to do something much longer and more in-depth than the typical journalistic piece (and rather eye-opening to contend with the timelines of academia first-hand).

SÓSeán Ó hÉigeartaigh@S_OhEigeartaigh

The 'AI race with China' was just used as justification to postpone the US Executive Order on AI safety and oversight, apparently after industry pressure. Makes this a good time to announce that my latest paper, with @yilingliu95 , @SnellColeman and @ShakeelHashim, has been accepted in the Cambridge Forum on Technology and Global Affairs. "When the Chips are Down: Technology Actors as Power Brokers in the US-China 'AI Race'" traces in depth the evolution of the 'AI race narrative from 2017 through to 2025, including who has been pushing it, and to what ends it has been deployed. While it has often looked organic, it has been heavily pushed, often in quite a coordinated way, by US tech industry actors - company CEOs, investors, lobby groups. It has gone through different phases under different administrations - the race to AGI, the diffusion race. A common factor is that (with a few notable exceptions, e.g. Anthropic) it has predominantly been used to promote interventions that financially benefit companies and their investors: federal R&D investment; adoption of their tech in government; laxer environmental laws; infrastructure buildout; laxer copyright laws; preventing antitrust action; even being used to undermine child safety protection. The race narrative has been weaponised with brutal effectiveness, despite how little purchase it has with academic experts or the American public. Often, as this week, safety and oversight has been the target. We bring receipts; lots and lots of receipts. Most of you know this has been a focus with me for a while. The reason is that I think it has created a distorted information/perception environment that has hugely undermined prospects for essential safety and governance (much of which wouldn't even undermine American competitiveness anyway - as long as they're selling the chips, that outweighs the impact of almost any safety intervention by a lot). There is of course good reason to take competitiveness seriously, and for the US to take the challenge from its nearest competitor, China, seriously. However, that is not incompatible with safety and oversight, or even carefully-scoped cooperation with China on mitigating risks, and this will soon be extremely necessary IMO. We need to be able to examine this in a clear-eyed way, and as a challenge to the dominant industry narrative, this paper is an attempt to help make this happen. Real privilege to write this with Coleman, and two brilliant AI journalists in Shakeel and Yi-Ling (Yi-Ling's own Transformer article in the comments is a companion piece to it). There's a lot in it; even if you think you know it all, I hope you might come across something interesting in there. Would love to hear thoughts. Preprint here: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6734224