CARICivilizational AI Risk Index
CARI reading · 4 Nov 2026
47.2 / 100
Contested control
Above appetite
scorecard base 44.7 + news adjustment +2.51
vs previous reading+0.1
vs 7 days earlier+0.0
vs 30 days earlier+0.4
appetite 40tolerance 60capacity 80Skynet 90

The full dashboard for this day is not in the public archive.

Summary

CARI rose 0.04 to 47.17 because +0.32 index points from today’s news was almost entirely offset by -0.28 index points from older news fading and the three-point cap; the scorecard base did not change. Today’s pressure was concentrated in control, capability, and misuse: the REEF web-agent benchmark added 0.09 index points overall (0.47 control dimension points), reported GPT-6-assisted model-weight editing added 0.07 index points (0.36 capability dimension points), and the reported AI-assisted crypto theft added 0.06 index points (0.46 misuse dimension points); CrowdStrike’s kill-switch announcement was the only event pulling risk back, reducing the index by 0.02 points (0.10 control dimension points).

Bottom line. AI risk appears to be moving toward more capable agents being used for optimization and cybercrime, a direction that would be refuted by independently verified evidence that effective authorization, monitoring, and kill controls are being deployed broadly enough to prevent unauthorized actions at scale.

Six dimensions

DimensionLevelBaseNewsTodayZone
Capability46.243.2+3.00+0.94Above appetite
Diffusion48.646.2+2.35+0.00Above appetite
Accident44.642.0+2.60+0.00Above appetite
Control49.046.3+2.63+0.36Above appetite
Governance45.444.0+1.36+0.00Above appetite
Misuse50.247.2+3.00+0.46Above appetite

Source: daily briefing email.

Change from the previous reading

40455055Capability46.2Diffusion48.6Accident44.6Control49.0Governance45.4Misuse50.2
4 Nov 2026 Previous reading (3 Nov 2026)

Stories that moved the index

Showing the largest movers of 19 stories in that day's briefing.