Who sets the rules for AI across borders, and how these bodies connect.
Point at a body to see its profile. Click a body, a line or a wedge to keep it. Press Esc to clear.
| From | Link | To | Since | Status | Confidence | Check | Sources |
|---|
A link means a source documents a specific relationship. Two bodies that only work on the same topic have no link. Solid lines are formal links. Dashed lines are informal links.
"Names in Official Text" is the weakest type, so the map hides it until you turn it on under Filters.
Some commitments belong to a country, not to one of its agencies (for example, a signature on a treaty). The map puts such a link on that country's node: its law or its agency. The link text says who acted.
AI research passes built the data. Then separate AI passes opened each source again and checked each link and each profile fact. Juan Guillen reviews the result. Each link shows the result: "Checked" or "Checked and Corrected". Facts that no source supports are left out.
Starting point: a diagram of international AI governance shared in an AI Safety chat in October 2026. I could not find its author; if it is yours, write to me and I will credit you.
Pairs you might expect to be linked, where the evidence says otherwise.
Questions open as of 24 October 2026.
Each number in the map, the profiles and the register points to one entry here. Primary sources are official texts or the body's own pages.
Research draft for juanguillen.me. Based on public sources opened between 24 and 25 October 2026. Links and status can change; check the sources before you cite.