✨🏛️ The political economy of AI: A Quick Q&A with … policy analyst Anton Leicht
'The perceived threat of an opponent gaining technological superiority soon has often induced conflict; I think this could be no different.'
While the fear of supersmart AI is common in today’s tech conversations, human politics may pose the more immediate danger. AI is becoming entangled in partisan struggles and international power competitions.
The current state of play, according to Anton Leicht in his article, The AI Takeoff Political Economy:
I feel that the path ahead is a lot more politically treacherous than most observers give it credit for . . . We need substantial changes to the political framing, coalition-building, and genuine policy planning around the ‘AGI transition’ – not (only) on narrow normative grounds. Otherwise, the chaos, volatility and conflict that can arise from messing up the political economy of the upcoming takeoff hurt everyone, whether you’re deal in risks, racing, or rapture. (Emphasis in original)
I asked Leicht a few quick questions about the politicization of AI and the risks surrounding AI proliferation, from state-based conflict to rogue terror threats.
Leicht is a policy specialist at KIRA, an independent think tank dedicated to facilitating AI while mitigating risk. He previously worked as a policy analyst and lobbyist for Berlin Global Advisors and served as a policy advisor in the Bundestag, the German Federal Parliament. He is currently pursuing a doctorate focused on the democratic governance of AI at Bayreuth University. Leicht is the author of his own Substack, Threading the Needle.
If there’s major job displacements, or an instance of AI-driven harm, it would be easy for Democrats to pin that on the administration’s light touch on and close ties to tech. That could become a winning political message — and if it does, I’d expect the shift back to come with very restrictive policy.
1) Now that tech is leaning to the right, how strong do you feel that association really is, and what would a sudden leftward shift of the political landscape look like for AI development and regulation?
I think the tech right alignment is strong for now. But it’s often born from backlash and frustration, and tech has always had a lot of turnover, both in terms of people and ideas — so I don’t think the association is certain to last. If politics shift back leftward again, effects on that alignment and on AI policy in general depend on how salient tech and AI is throughout the current administration. Currently, it looks like the administration’s tech policy is not the most politically visible or controversial part of its agenda, so it might just avoid the backlash.
That could change if politically salient AI-related harms manifest in the next few years. If there’s major job displacements, or an instance of AI-driven harm, it would be easy for Democrats to pin that on the administration’s light touch on and close ties to tech. That could become a winning political message — and if it does, I’d expect the shift back to come with very restrictive policy. The exact content depends on the kind of harm: If it’s jobs, the crackdown might be on commercial deployment of AI; if it’s misuse, it might be through extensive pre-deployment oversight or liability. But either way, if the Democrats overplay their hand on the rhetoric and substance around that, it’s easy to create the impression of ‘persecuting’ tech — and that might be all it takes for the tech right association to actually stick.
. . . I’m very confident there will still be policy to be made in 2028 and 2032, and overcommitting to the current incidental alignment could turn out to be a mistake.
2) How do we decouple AI advancement from partisanship?
I think both sides and the AI developers have a part to play here.
On the Republican side, avoiding an open political flank is important: Make some effort to prevent harm from and fear of AI on your watch, so that you don’t give Democrats too much of a political incentive to turn to tech crackdowns.
On the Democrat side, it’s trying to avoid giving into the temptation of ‘going after’ the tech industry in response to 2024. We saw a micro version of that with the past administration’s treatment of cryptocurrency which, no matter the policy merits, has politically galvanised part of the tech ecosystem.
And on the developer side, keeping a healthy distance would be very productive. It’s tempting to go all-in on alignment with the GOP and its rhetoric, especially if you think the next few years are very important. But I’m very confident there will still be policy to be made in 2028 and 2032, and overcommitting to the current incidental alignment could turn out to be a mistake.
