About AI Policy Watch

AI Policy Watch is a public-interest aggregator of current federal government activity related to artificial intelligence. It exists so that members of the public can see what their government is deciding about AI, form a view, and participate in the process if they choose to.

What this site tracks

Three kinds of federal activity, each linking out to the authoritative source rather than hosting its own detail pages.

• Legislation — federal bills related to AI, in progress or resolved. Source: Congress.gov.
• RFIs — federal Requests for Information and similar public-input opportunities. Source: Federal Register, with Regulations.gov links where available.
• Hearings — congressional committee meetings where AI is on the agenda. Source: Congress.gov. Coverage is partial; committees publish their own schedules and not all appear in the API.

How we decide what counts as AI

A bill, RFI, or hearing is included if either:

• It carries the Library of Congress's "Artificial Intelligence" subject tag (primary filter, bills only); or
• Its title or abstract contains a term from a deliberately conservative keyword list — artificial intelligence, generative AI, machine learning, algorithmic accountability, automated decision-making, facial recognition, deepfake, synthetic media, large language model, and a handful of adjacent terms. The full list is at config/keywords.yml and can be proposed for change via pull request.

When in doubt we exclude rather than include. A manual include/exclude list (in the repository) lets us correct specific edge cases and is itself open to contribution.

What this site doesn't do

• Take a position on AI policy. We present information; you decide.
• Replace enterprise trackers. If you're a lawyer, compliance officer, or government-affairs professional, tools like MultiState, IAPP, and Steptoe are built for your needs. This one is built for the general public.
• Cover state-level or international AI policy. V1 is federal-only. State coverage is planned for V2.
• Host editorial summaries. V1 mirrors source-language titles. Plain-language summaries are planned for V2.
• Offer user accounts, saved searches, or email alerts.

How often data updates

Daily sync via GitHub Actions at 10:00 UTC (about 6:00 AM Eastern). Tunable as we see API behavior in practice. The dataset was most recently refreshed on May 15, 2026.

How to contribute

The project is open source. Corrections to the dataset, proposals to adjust the keyword list, copy fixes, bug reports, and code contributions are all welcome via GitHub.

• Repository
• Contributing guide
• File an issue

A non-technical feedback path (in-site forms that open issues) is planned as the first fast-follow.

Who runs this

AI Policy Watch is maintained by Natasha Najdovski, with the intention of adding co-maintainers as the project matures. It is independent and non-commercial.

The project is built with substantial help from AI coding agents (primarily Claude, from Anthropic), under human review. We disclose this because it would be inconsistent not to: the mission of this site is to make federal AI policy visible, and that includes being visible about how this particular AI-adjacent tool is itself built. The rules agents follow are documented in AGENTS.md; agent-assisted commits are tagged in the git history.

How to give feedback

The best path today is a GitHub issue. Bugs, missing items, mis-included items, suggested keywords, and copy corrections are all welcome.