WORK
Weidinger PNAS Study
The 2023 Proceedings of the National Academy of Sciences paper by Laura Weidinger and colleagues at Google DeepMind that operationalized the veil of ignorance as an experimental protocol — and found that participants behind the veil reliably chose principles prioritizing the least advantaged.
Across five incentive-compatible studies with over 2,500 participants, Weidinger and colleagues asked people to choose principles to govern an AI assistant. Some participants chose from behind a simulated veil — without knowledge of their relative position in the group. Others chose with full knowledge of their position. The result was consistent and robust across study variations: participants behind the veil showed a clear preference for principles instructing the AI to prioritize those worst-off. Neither risk attitudes nor political preferences adequately explained these choices. The preference appeared to be driven by elevated concerns about fairness — precisely what Rawls's framework predicts. The study is methodologically significant because it transformed the veil of ignorance from a philosopher's thought experiment into an empirically validated mechanism for aligning AI systems with principles of justice.
In The You On AI Field Guide
The study's design was elegant. Participants were placed in small groups