Who makes the rules for AI?
This session is part of the How to Think about Tech? The Case of ‘AI Safety’ study group* initiated by some of the fellow candidates of the 2025/2026 Introduction to Political Technology course. It is open to faculty and fellowship candidates only.
Despite the more restricted sense of governmental control of economic behaviour, regulation is a broader matter of rules that can come from various sources, take various forms, and operate at various levels. This session examines AI regulation as a socio-technical and political process, analysing how governance emerges through a fragmented ecology of laws, codes, expert discourse, corporate influence and civil society resistance. By reading regulatory texts alongside critiques of economies of influence and industry-led governance, we will explore how AI regulation is enacted, negotiated, and contested across global and national arenas.
Recommended readings:
Approaches to regulation
Michael Vale et. al - AI and Global Governance: Modalities, Rationales, Tensions
Julia Black - Decentred regulation
Lawrence Lessing - Code Is Law
Ada Lovelace Institute - Safe before sale
Economies of influence
Lawrence Lessing- Protected Democracy
Article 19 - EU: Code of Practice on General Purpose AI – a masterclass in corporate compromise
Minwoo Ki et al - Project 14: Beyond Code - The Battle for Influence in AI's Global Rulebook or Civil society open letter demands to ensure fundamental rights protections in the Council position on the AI Act - AlgorithmWatch
Kewin Wei at al - How Do AI Companies "Fine-Tune" Policy? Examining Regulatory Capture in AI Governance
Katie Mcque et al - The global struggle over how to regulate AI - Rest of World
Regulatory acts
EU AI Act or EU AI Act Chapter 5 or EU General-Purpose AI Code of Practice
California Legislative SB-53 Artificial intelligence models: large developers
Ensuring a National Policy Framework for Artificial Intelligence – The White House
Ada Lovelace Institute - Will the UK AI Bill protect people and society?
Podcasts:
Dean Ball on how AI is a huge deal — but we shouldn't regulate it yet | 80,000 Hours
Markus Anderljung on how to regulate cutting-edge AI models - 80,000 Hours
Something to have fun with:
You are invited to analyse EU consultation as a governance instrument, i.e. a device that shapes who can speak, what counts as expertise, and which futures for AI become legible to regulators. Review the Commission's draft on establishing AI regulatory sandboxes by focusing on: Who is this consultation written for? What problems does the consultation make visible and what does it obscure? How is ‘innovation’ balanced against protection? What kinds of influence does the format invite? What would meaningful feedback look like from a public-interest perspective?
To have a fruitful discussion, please read at least one article from each main category of readings.