Analysis
Seven Findings
Computed from 364 policies across 133 governments, 2000–2023 · Harvard Dataverse V1, July 2026
Adoption clusters into a single decade.
73% of all policies were enacted after 2015 — the field barely existed before the last ten years.
Open the chartContested terms, rarely defined.
Of policies invoking “disinformation,” only 28% actually define it. The label does the work the definition should.
Open the chartA handful of platforms absorb the attention.
Naming Facebook, X, and YouTube is the modal regulatory move — the long tail of services is rarely touched.
Open the chartCoverage is wide but shallow.
133 governments appear, yet 21% of them contribute just a single policy.
Open the chartThe targeted technology shifts under the policy.
Attention moves from broadcast and print toward social platforms — and, most recently, toward AI.
Open the chartHow governments punish depends on where they are.
Criminal, civil, and administrative liability split sharply by region — the same conduct carries very different stakes.
Open the chartA small set of states recur as named threats.
Russia, Iran, and China dominate the named-actor network — most other states appear once, if at all.
Open the chart