Stanford AI Tool Scans 500M Words to Cut Government Red Tape

Stanford researchers have created an artificial intelligence system that scanned 500 million words of state statutes to identify outdated reporting requirements, commissions, and fees — and they're now working directly with state governments to eliminate the bureaucratic clutter.

The tool, developed by scholars at Stanford's Institute for Human-Centered AI (HAI) and Stanford RegLab, examined legal code across all 50 U.S. states to reveal patterns in what experts call "policy sludge" — obsolete provisions that trap civil servants in administrative burdens and slow government operations.

Why it matters

Millions of Americans encounter government delays when starting businesses, obtaining licenses, or navigating public services. The research demonstrates that much of this friction stems from requirements nobody reads or completes. By quantifying the problem and providing tools to address it, the work offers a concrete path to making government more responsive without requiring new legislation or funding — just systematic cleanup of existing rules.

Measuring the bloat

The scale of the problem surprised even the researchers. In California, reporting requirements grew 400 percent between 2000 and 2025. Maryland's backlog became so severe that reading all mandated reports would take 14 weeks — longer than the state's 13-week legislative session.

Perhaps most striking: 30 percent of California's ongoing reports may never have been completed at all. When Maryland agencies reviewed their own requirements, they flagged 20 percent as candidates for elimination or consolidation.

The team, led by Stanford Law professor Daniel E. Ho, published their findings in a forthcoming paper titled "The Abundance of Reports and Incapacity of States" in the Yale Journal on Regulation. Coauthors include RegLab researchers Emily Robitschek, Ananya Karthik, Gabe Malek, and Derek Ouyang.

Real-world impact

New York Governor Kathy Hochul issued an executive order this month directing state agencies to conduct a "regulatory reset" based on the research. The order mandates removal of outdated requirements, burdensome fees, and obsolete reports and commissions.

California is using the data to convert paper reports into digital dashboards. San Francisco previously used the tool to streamline over one-third of the city's reporting requirements through legislation.

"The AI tool built by the RegLab team was instrumental in New York's Regulatory Reset, enabling us to convert unwieldy legalese into digestible datasets," said Zoe Jacobs, director of regulatory reform in Governor Hochul's office.

Cost disparities

The research revealed dramatic variation in reporting burden. One single report consumed 3,500 staff hours and over $870,000 to produce. Others required only a few hours and were viewed thousands of times, demonstrating genuine value.

The team also uncovered absurd relics. New York still maintains a requirement that the Board of Regents report on actions against "subversive" teachers — a Red Scare measure the Supreme Court found unconstitutional in 1967.

Tools for reform

RegLab released the full scan results across all 50 states on a public website, allowing any jurisdiction to explore reporting requirements. The team also developed a model state statute that includes provisions for automatic sunsetting of requirements, digital repositories, and lightweight cost-benefit tracking.

The research tested political assumptions about government bloat. While reporting requirements proved more prevalent in Democratic states, the partisan difference was small relative to the overall scale — the problem exists across party lines.

These details were first reported by Stanford HAI.