AI scan of 24 years of Singapore property deals exposes officials' insider metro-station profits

_akpiper · x · 2026-09-23

A widely discussed NBER paper had Stanford and Columbia researchers use AI to scan all Singapore property transactions from 1995–2019, cross-referenced with a registry of 141,000 civil servants. Officials bought property near future metro stations 1–2 years before official announcements at a 60% higher-than-normal rate, reselling at 12% annualized returns; relatives showed the same pattern, totaling S$270M. Real-estate brokers and company directors showed no such pattern, and 1,000 random simulations confirmed the effect — a landmark case of AI-powered corruption auditing.

Original post →

More from AGI Musings

AGI Musings channel →