These visualizations are experimental and still being refined. Designs and data presentation may change.

Governance Gap by Risk and Sector

Every risk subdomain against every economic sector. Each cell compares how vulnerable experts judge that sector to be to that risk against how well the governance corpus actually covers the pairing. Click any cell to see the documents and arithmetic behind it.

Each cell is a signed gap: the expert vulnerability rating for that sector and risk (normalized to 0–1) minus how well the governance corpus covers the pairing. Positive (red) means vulnerability exceeds coverage — the pairing is under-governed. Negative (blue) is the reverse: governance coverage runs ahead of the assessed vulnerability. The color scale is symmetric around zero and derived from the data, so it rescales when you switch metrics.

The two metrics differ in what counts as coverage. Breadth & depth combines how many of the documents covering a risk also address the sector with how substantively they treat the risk (a floor is applied, so even a superficial mention counts as baseline coverage). Breadth only drops the depth term and counts mentions alone, which makes more cells negative.

One consequence is worth naming: a risk that no document in the corpus mentions has zero coverage by construction, so its entire row scores the full vulnerability rating rather than something neutral. Subdomain 7.5 (AI welfare) sits in exactly that position in the U.S. federal corpus.

	Agri/Mining/Constr/Mfg Trade/Transport/Utilities Information Finance & Insurance Real Estate Prof & Technical Svcs Scientific R&D Mgmt/Admin/Support Educational Services Health Care & Social Arts/Entertainment Hospitality & Other Public Admin (excl NatSec) National Security
1.1 Discrimination	0.15	0.06	0.48	0.52	0.30	0.28	0.22	0.39	0.51	0.41	0.46	0.13	0.14	0.26
1.2 Toxic content	0.18	0.32	0.34	0.45	0.07	0.09	0.14	0.63	0.64	0.58	0.50	0.45	0.18	0.57
1.3 Unequal performance	0.21	0.12	0.26	0.57	0.35	0.07	-0.02	0.45	0.57	0.45	0.24	0.18	0.20	0.44
2.1 Loss of privacy	0.22	0.16	0.50	0.64	0.48	0.34	0.30	0.25	0.34	0.57	0.20	0.24	0.52	0.58
2.2 AI security vulnerabilities	0.22	0.45	0.58	0.73	0.57	0.40	0.27	0.60	0.40	0.67	0.34	0.36	0.33	0.50
3.1 False information	0.41	0.34	0.48	0.76	0.34	0.34	0.41	0.42	0.62	0.68	0.24	0.41	0.28	0.58
3.2 Loss of consensus reality	0.45	0.32	0.43	0.46	0.34	0.33	0.34	0.50	0.37	0.71	0.17	0.45	0.23	0.55
4.1 Disinformation & influence	0.19	0.26	0.39	0.36	0.26	0.29	0.26	0.40	0.42	0.38	0.27	0.44	0.21	0.45
4.2 AI weapons & cyberattacks	0.11	0.39	0.45	0.65	0.30	0.34	0.15	0.35	0.39	0.62	0.36	0.40	0.53	0.35
4.3 AI fraud & scams	0.31	0.53	0.37	0.60	0.52	0.25	0.37	0.59	0.34	0.57	0.41	0.34	0.19	0.31
5.1 Overreliance & unsafe use	0.26	0.23	0.53	0.65	0.24	0.42	0.30	0.55	0.60	0.48	0.25	0.27	0.28	0.67
5.2 Loss of human agency	0.20	0.39	0.53	0.61	0.40	0.34	0.30	0.45	0.57	0.52	0.14	0.15	0.28	0.62
6.1 Power centralization	0.29	0.60	0.72	0.83	0.36	0.43	0.40	0.35	0.43	0.57	0.63	0.35	0.42	0.68
6.2 Inequality & unemployment	0.09	0.11	0.51	0.38	0.18	0.30	0.29	0.61	0.50	0.36	0.44	0.14	0.23	0.14
6.3 Devaluation of human creativity	0.39	0.39	0.70	0.61	0.36	0.51	0.51	0.64	0.75	0.59	0.76	0.39	0.24	0.31
6.4 Competitive dynamics	0.16	0.50	0.58	0.49	0.30	0.40	0.30	0.37	0.48	0.49	0.65	0.14	0.44	0.54
6.5 Governance failure	0.16	0.43	0.49	0.65	0.24	0.33	0.21	0.53	0.36	0.60	0.55	0.30	0.21	0.44
6.6 Environmental harm	0.69	0.66	0.56	0.80	0.40	0.32	0.38	0.41	0.18	0.76	0.41	0.65	0.39	0.68
7.1 AI misalignment	0.39	0.41	0.63	0.79	0.32	0.43	0.36	0.68	0.35	0.36	0.45	0.45	0.31	0.45
7.2 Dangerous capabilities	0.15	0.50	0.49	0.67	0.27	0.39	0.19	0.39	0.42	0.39	0.37	0.46	0.33	0.39
7.3 Capability & robustness	0.25	0.43	0.58	0.74	0.28	0.41	0.25	0.33	0.37	0.59	0.35	0.32	0.27	0.54
7.4 Transparency & interpretability	0.19	0.14	0.44	0.59	0.15	0.32	0.19	0.23	0.29	0.48	0.20	0.21	0.45	0.52
7.6 Multi-agent risks	0.12	0.40	0.67	0.63	0.17	0.38	0.34	0.17	0.37	0.38	0.17	0.17	0.60	0.72

← governance exceeds vulnerability-0.83+0.83vulnerability exceeds governance →

Key Takeaways

1.6.1 Power centralization × Finance and Insurance is the widest single gap (0.83) — 29% of the documents covering that risk mention the sector, and 0 of 6 cover the risk well.
2.Only 1 of 336 cells are negative — governance exceeds assessed vulnerability almost nowhere.
3.Finance and Insurance is the widest column — 6 of the 10 largest gaps sit in it.
4.7.5 AI welfare is mentioned by no document in this corpus, so its row scores the full vulnerability rating by construction.

Vulnerability ratings come from the Delphi expert survey; coverage is computed from the governance corpus. Cells resting on very few documents are flagged in their detail view.