A structured, queryable database tracking AI laws, regulations, and governance frameworks worldwide.

🚀 Use this tool live — No installation required. Search and filter legislation in your browser.

Data Currency • What's Different • Quick Reference • Getting Started • Use the Data • Contribute

AI governance is fragmenting fast, and it does not hold still. The EU AI Act's high-risk obligations were pushed from 2026 to 2027. Colorado's law — the first comprehensive US state AI statute — was delayed, enforcement-stayed, then repealed and replaced before it ever took effect. Canada's AIDA died at prorogation. Keeping track requires structured data with a history, not news alerts.

This repo provides:

• Machine-readable JSON datasets with a consistent schema and permanent record ids
• Supersession tracking — what replaced what, and when
• CLI tools for querying across jurisdictions
• A browser interface with no install and no account
• Zero dependencies — pure Python standard library

Last review: September 2026. 44 entries across 22 jurisdictions.

Each entry carries its own last_verified date and a verification field saying how it was checked:

This is a research and tracking tool, not legal advice. Always confirm against the official source before making a compliance decision. Every entry links to one.

Most AI law trackers are web pages or PDFs published by law firms and membership bodies. They tell you what a law says today. They are excellent, and they are not reusable — you cannot query them, diff them, or cite a stable identifier.

This project is built around two things those cannot offer:

1. The data is yours. MIT licensed, plain JSON, fetchable from a stable URL. Load it into a notebook, a spreadsheet, a compliance tool or a retrieval index without asking anyone.

2. Records have a history. Ids are permanent. state-001 means Colorado SB 24-205 forever, including now that it has been repealed — the record stays, its status becomes superseded, and superseded_by points at the statute that replaced it. A citation written in 2024 still resolves, and tells you what changed underneath it. Every change is a git commit with a source link.

This dataset deliberately tracks notable frameworks and instruments with operative legal effect, not every introduced bill. Depth and accuracy over volume.

Run python3 src/generate_dashboard.py for the current version of this table, generated from the data.

git clone https://github.com/delschlangen/ai-legislation-tracker.git
cd ai-legislation-tracker

# Full dashboard report
python3 src/generate_dashboard.py

# Query by topic, status or jurisdiction
python3 src/query_legislation.py --tag employment
python3 src/query_legislation.py --status superseded
python3 src/query_legislation.py --jurisdiction Colorado
python3 src/query_legislation.py --search "frontier"
python3 src/query_legislation.py --list-tags

# Check the data is well formed
python3 validate.py

# Export everything to one CSV for a spreadsheet
python3 src/export_csv.py

Requirements: Python 3.7+. No external dependencies.

The three JSON files are the whole dataset and are stable URLs:

https://raw.githubusercontent.com/delschlangen/ai-legislation-tracker/main/data/us_federal_actions.json
https://raw.githubusercontent.com/delschlangen/ai-legislation-tracker/main/data/us_state_bills.json
https://raw.githubusercontent.com/delschlangen/ai-legislation-tracker/main/data/international_frameworks.json

import json, urllib.request

BASE = "https://raw.githubusercontent.com/delschlangen/ai-legislation-tracker/main/data"
state = json.load(urllib.request.urlopen(f"{BASE}/us_state_bills.json"))

for law in state:
    if law["status"] == "enacted":
        print(law["bill_number"], "-", law["title"])

Prefer a spreadsheet? python3 src/export_csv.py flattens all three files into one CSV. It is generated on demand rather than committed, so there is never a stale second copy of the data in the repository.

{
  "id": "state-013",
  "state": "California",
  "bill_number": "SB 53",
  "title": "Transparency in Frontier Artificial Intelligence Act",
  "status": "enacted",
  "date_enacted": "2025-09-29",
  "effective_date": "2026-01-01",
  "summary": "First US state frontier AI safety law...",
  "key_provisions": ["Publish a frontier AI framework...", "..."],
  "source_url": "https://leginfo.legislature.ca.gov/...",
  "tags": ["frontier_ai", "safety", "transparency"],
  "last_verified": "2026-09-17",
  "verification": "secondary"
}

ai-legislation-tracker/
├── data/                            # The dataset — the single source of truth
│   ├── us_federal_actions.json
│   ├── us_state_bills.json
│   └── international_frameworks.json
├── docs/                            # GitHub Pages site; reads data/ at runtime
│   ├── index.html
│   ├── app.js
│   └── styles.css
├── src/
│   ├── generate_dashboard.py        # Markdown summary report
│   ├── query_legislation.py         # CLI filtering and search
│   └── export_csv.py                # Flatten the dataset to CSV on demand
├── .github/
│   ├── ISSUE_TEMPLATE/              # Forms for reporting or suggesting entries
│   └── workflows/validate.yml       # Read-only data check on pull requests
├── examples/current_landscape.md    # Sample generated dashboard
├── validate.py                      # Data validator (read-only)
├── CITATION.md · CONTRIBUTING.md · MAINTENANCE.md · LICENSE
└── README.md

You do not need to know Git. If you spot something wrong, open an issue:

• Report an incorrect or out-of-date entry
• Suggest a law to add

Corrections are the most valuable contribution here. A tracker that is confidently wrong is worse than one that is visibly incomplete.

If you would rather edit the data directly, see CONTRIBUTING.md for the schema and the submission checklist. Run python3 validate.py before opening a pull request — the same check runs automatically and will tell you what is wrong.

• Track supersession and repeal, not just current status
• Single source of truth for the website and the dataset
• Read-only validation on pull requests
• Issue forms for non-technical contributors
• CSV export for spreadsheet users
• Re-verify every entry against primary sources and move verification to primary
• Publish a JSON Schema file
• Broaden state coverage, prioritising laws with operative obligations

MIT License — See LICENSE

Data is curated from public government sources.

Built by Del Schlangen
LinkedIn