Work that takes the discourse around AI — how firms describe their own conduct, which coalitions form around policy questions, and how framing constrains the range of thinkable regulation — as the object of analysis rather than as commentary on it.
It is worth separating from technical and legal analysis because it explains outcomes those cannot. Whether a bill passes often turns less on its merits than on which coalitions assemble around it and whose framing of the question prevails.
Two lines of work
The stated-principles gap. Reisner's Atlantic feature documents an intellectual-property double standard: AI companies "aggressively defend their own patents, terms-of-service clauses, and software piracy protections while claiming 'fair use' over copyrighted books, videos, and music to train their models." The analytical move is comparative rather than accusatory — it holds a firm's conduct in one domain against its conduct in another, which is a form of evidence that neither a technical assessment nor a legal brief supplies.
Coalition formation. Transformer's reporting on the defeat of the proposed ten-year state AI-regulation moratorium in summer 2025 traces the bipartisan coalition that assembled against it — Marjorie Taylor Greene, Steve Bannon, Josh Hawley, Sarah Huckabee Sanders, evangelical pastors, and labor unions — around opposition to tech-industry preemption. The finding is that AI policy alignments do not map onto conventional partisan axes, which is why forecasts assuming they do have repeatedly mispredicted outcomes. See State-Level AI Regulation.
Why the category is useful
Both lines identify a mechanism that determines policy outcomes without appearing in the policy text. The first concerns credibility: a firm's advocacy carries different weight once its conduct in adjacent domains is on the record. The second concerns coalitions: preemption failed not because the substantive arguments changed but because an unusual alliance formed. Neither is visible from the instruments themselves.
The category also covers the framing question that Regulatory Capture in AI Policy addresses from the political-economy side — how far the terms of debate are set by the firms being debated.
Relationships
- related: Regulatory Capture in AI Policy — the political-economy account of the same dynamic
- related: The Hypocrisy at the Heart of the AI Industry — Alex Reisner (The Atlantic, March 2026), How MAGA learned to love AI safety — Nicky Woolf (Transformer, October 2025), State-Level AI Regulation, AI Copyright, AI Policy (umbrella)