US AI Policy Portal — Case study

What this project demonstrates

The skills an AI governance hire actually needs — applied end-to-end on a real problem, not as line items on a résumé.

Policy synthesis across 50+ jurisdictions and constantly-shifting source material
Original taxonomy design (comprehensive vs. harm-reactive framework)
Technical execution — front-end build, data structure, deployment
UX judgment for a research audience (compliance teams, journalists)
Self-directed scope and ongoing maintenance discipline
Public communication of complex regulatory material

The question

State AI legislation has accelerated dramatically over the past 24 months. The federal landscape remains contested. The result is a patchwork that's hard to read at a glance: which states have done what, how their approaches differ, what's coming into effect when, and which provisions are doing real regulatory work versus signaling political posture.

Researchers, compliance teams, and journalists kept asking the same comparison questions. I built the Portal to answer them with primary-source data instead of vibes.

What it does

State-by-state browse

Click any state to see its enacted AI laws, scope, regulated actors, and effective dates.

Three-state comparison

Side-by-side view of approach, scope, and timing across any combination of three states.

Filter and slice

By regulatory approach, harm category, year of effect, scope, or regulated actor.

CSV export

Pull the underlying data for your own analysis or citation in policy briefs and reporting.

What I found

states with comprehensive frameworks: California, Colorado, New York, Utah, Texas

75%

of provisions are harm-reactive (CSAM, deepfakes, fraud) rather than systemic

2026

the implementation wave most institutions aren't ready for

The picture that emerges: American AI regulation is forming bottom-up, state-by-state, and disproportionately around discrete harms rather than systemic governance of AI development itself. Five states are doing something structurally different — and the gap between them and everyone else is the story to watch over the next eighteen months.

Who it's for

Policy researchers tracking regulatory trends across jurisdictions
Compliance and legal teams scoping multi-state AI deployments
Journalists looking to cite or compare state approaches with primary sources
Anyone trying to understand where American AI regulation is actually heading

Methodology

Legislation sourced from state legislatures, NCSL, and MultiState's AI tracker. Each law is coded across structural dimensions — scope, regulated actors, enforcement mechanism, harm category, effective date. The comprehensive vs. harm-reactive classification follows a taxonomy I'm refining publicly through the Substack series. The Portal is updated as new laws are enacted; the underlying data is available via CSV export.