AI Policy Wiki
Dashboard

Inside My Advanced Topics Class 8.2: The Environment Is the Argument (Farahany, March 2026)

medium confidence · updated 2026-06-06

The architecture-vs-information theory of consent. Even if users read every word, the choice environment (virtual currencies degrading price salience, purchase buttons next to preview buttons, daily-rotating shops manufacturing urgency) is engineered to produce particular outcomes regardless of disclosure. Anchored on Fortnite's V-Bucks system and the FTC's $245M Epic settlement.

Author: Nita Farahany Source: https://nitafarahany.substack.com/p/the-environment-is-the-argument-inside Published: March 19, 2026

This is the second installment of Week 8 in Nita Farahany's Advanced Topics in AI Law and Policy course (Class 8.2 of approximately 30). The class advances an architectural critique of consent: even when every disclosure is perfect and every user reads every word, the environment in which consent is exercised is itself engineered to produce particular outcomes. The argument is that disclosure-theory law cannot reach environmental architecture.

Summary of argument

The class distinguishes two theories of consent. Under the information theory, the legal question is whether the user received and understood the relevant information; under the architecture theory, the question is whether the choice environment was engineered to steer the user toward a particular result regardless of what they understood. Farahany argues that the five forms of architectural manipulation described below operate by exploiting attention, time pressure, and cognitive load, so that information which would be easy to process under reflective conditions becomes harder to process under engineered conditions. On this account, disclosure remedies the information layer while architecture exploits the conditions under which information is processed, leaving disclosure remedies unable to fix architecture.

The information-theory-versus-architecture-theory split is presented as the through-line of Weeks 8 and 9. Farahany argues that laws built on the information theory (GDPR notice-and-consent, COPPA disclosure, CCPA) leave architecture untouched, while laws built on the architecture theory (the UK Children's Code, CAADCA, and the FTC click-to-cancel rule) face First Amendment and vagueness challenges precisely because they regulate design.

Key claims

The class is anchored on Fortnite's V-Bucks system and the FTC's $245M Epic settlement (FTC v. Epic Games, 2022), which Farahany reads as cataloging several features of an engineered purchase environment: a virtual currency that degrades price salience, so that children pay "1,000 V-Bucks" rather than "$10"; purchase buttons placed adjacent to preview buttons, default-confirmed transactions, and saved-card defaults; daily rotating shops with countdown timers manufacturing urgency; and difficulty-of-refund engineering after the purchase.

From the FTC dark-patterns report and the EU DSA implementation acts, the class synthesizes five categories of architectural manipulation:

  1. Salience manipulation — making the cost less perceptually obvious, through virtual currencies, fine print, color, or font size.
  2. Friction asymmetry — purchase is one click, while refund or cancellation requires multiple steps.
  3. Default engineering — the default option benefits the platform rather than the user.
  4. Urgency manufacture — countdown timers, scarcity claims, and social-proof claims that pressure rapid decisions.
  5. Choice architecture — the order, placement, and visual hierarchy of options is optimized for the platform's outcome.

The class sets up Class 8.3, which examines three legislative responses to the architecture problem (the DELETE Act, COPPA, and CAADC). Farahany frames these as a test of whether US law can move from information-theory remedies to architecture-theory remedies, and whether the constitutional system permits the move.

Relationships