Stanford Researchers Use AI to Map 500 Million Words of State Reporting Rules

07:15 · July 29, 2026

State legislatures have spent decades adding new reporting requirements onto agencies without ever repealing the old ones, and until now nobody had a clear picture of how large that pile had become. Stanford HAI and Stanford RegLab researchers used AI language tools to comb through roughly 500 million words of state statutes nationwide, tagging every provision that forces an agency to produce a report, and the results were striking. California’s reporting requirements grew 400 percent between 2000 and 2025, and the team estimates about 30 percent of the state’s ongoing reports may never actually be filed. In Maryland, agencies themselves identified 20 percent of their reports as candidates for elimination, and reading every report due in a single session would take 14 weeks, longer than Maryland’s 13-week legislative session allows. One report alone was calculated to consume 3,500 staff hours and more than $870,000 to produce. New York, meanwhile, still carries a 1960s-era statute requiring reports on ‘subversive’ teachers, a provision the Supreme Court invalidated in 1967 but that nobody ever bothered to strike from the books. The researchers are now working directly with several states to use the same AI-assisted method to identify and retire obsolete reporting rules.