From Privacy Washing to Sustainable Data Strategies: A Theory-Based AI Approach
(A2026-133671)

AUTHORS

Ziqi Zhong, University of Technology Sydney/The London School of Economics and Political Science

ABSTRACT

We develop analytical models characterizing when sustainable data strategies enhance profits and consumer surplus. To enable scalable measurement, we introduce PrivaAI, an AI agent fine-tuned on 28,360 human evaluations (1,418 policies) achieving 92.5% alignment with consumer perceptions. Analyzing policies from 10 million websites reveals: (i) systematic privacy deterioration with market position, (ii) substantial industry heterogeneity, (iii) regulatory asymmetries—GDPR yields balanced improvements while CCPA shows uneven patterns.A pre-registered experiment (N=610) confirms PrivaAI predicts consumer behavior and uncovers an "authenticity paradox": consumers perceive high-quality policies with dimensional inconsistency as less deceptive than uniform excellence, interpreting selective excellence as genuine prioritization rather than privacy washing. These findings provide a scalable foundation for benchmarking privacy practices with actionable insights for firms and regulators.