Measured global regulation

Where AI law exists—and where our knowledge stops.

A source-linked view of laws, regulations and enforcement. Empty cells remain empty; access barriers and our own unfinished work are never presented as the same fact.

Last measured30 Oct 20262-day median record freshness
1,517AI-specific instruments tracked127 jurisdictions · 100% source-linked225In forceNot proposed or merely enacted247Publish-ready enforcement matters39 jurisdictions203Active official-source readers22 stale · 23 faulted75Jurisdictions in measured scope195 historical-axis sweeps open · 4 unknown cells

Coverage, not a row-count heatmap

World coverage across three separate axes

CoveredOur gapAnchor onlyBlockedNot publishedNone foundDeferredUnknown
How to read an empty country

White does not mean “no AI law.” It means we make no claim because that cell was not measured or was deliberately deferred.

Grey is not our gap. The source is blocked or the authority does not publish; the country panel tells you which.

Amber is our gap. A readable source remains to be ingested, or only a general anchor law is held.

Enforcement volume

Recorded AI enforcement by year

2
2015
2
2016
3
2017
8
2018
4
2019
6
2020
18
2021
17
2022
21
2023
45
2024
60
2025
58
2026

The current year is incomplete. 3 publish-ready matters have no usable filed or decided date and are excluded from this chart.

What regulators act on

AI role in the matter

Biometrics & facial recognition75
Unspecified AI role68
Automated decisions40
Chatbot safety21
Deepfakes & synthetic media16
Training-data intellectual property14
Algorithmic pricing13
Deceptive AI claims13
Recommendation systems13

“Unspecified” stays visible: it means the source establishes an AI connection without naming the AI function.

United States

In-force law and proposals are not the same thing

29 jurisdictions with an in-force instrument20 with proposals but none in force

117 in-force state-level instruments and 856 pending or proposed instruments are counted separately.

Official sources297
Active readers203
Over silence budget22
Consecutive faults23
Snapshot size118 KB

Deliberately not shown

A dashboard is also defined by what it refuses to claim.

Penalty totals

Not shown until matter-level and multistate deduplication plus imposed dates are complete.

Upcoming effective dates

Not shown until the forward-looking rulemaking layer passes the AI-specific scope gate.

Verification-tier chart

Not charted while the public corpus has only a single-source verification tier.