The full dashboard for this day is not in the public archive.
Summary
CARI was unchanged at 47.29: against an unchanged 44.66 scorecard base, today’s news added +0.31 index points, exactly offset by -0.31 from older news fading and the 3.00-point cap, leaving a net news adjustment of 2.63. The upward pressure was led by the reported Pentagon personal-agent incident, which contributed +0.16 index points overall, +0.40 accident dimension points and +0.43 control dimension points, while North Korean threat-actor adoption added +0.06 index points overall and +0.46 misuse dimension points. OpenAI’s claimed misalignment monitoring was the clearest listed event-level pullback at -0.10 index points overall and -0.50 control dimension points, but older-news fading offset the aggregate move; all six dimensions remain above appetite, with misuse highest at 50.00.
Bottom line. AI risk appears to be shifting from capability growth toward operational-control failures and misuse; repeated unauthorized-agent access and verified threat-actor adoption would confirm that trajectory, while durable controls and declining misuse would refute it.
Six dimensions
| Dimension | Level | Base | News | Today | Zone |
|---|---|---|---|---|---|
| Capability | 46.2 | 43.2 | +3.00 | +0.59 | Above appetite |
| Diffusion | 49.2 | 46.2 | +2.93 | +0.21 | Above appetite |
| Accident | 45.0 | 42.0 | +3.00 | +0.65 | Above appetite |
| Control | 48.7 | 46.3 | +2.35 | −0.07 | Above appetite |
| Governance | 45.7 | 44.0 | +1.67 | +0.00 | Above appetite |
| Misuse | 50.0 | 47.2 | +2.82 | +0.46 | Above appetite |
Source: daily briefing email.
Change from the previous reading
Stories that moved the index
- A Defense Contractor Gave Their Personal AI Agent Access to a Secure Pentagon System+0.2 pts · Wired, Oct 29, 2026
- How we detect frontier model misalignment−0.1 pts · OpenAI Blog, Oct 29, 2026
- Tracking AI adoption in North Korean threat actor tradecraft+0.1 pts · Microsoft, Oct 29, 2026
Showing the largest movers of 23 stories in that day's briefing.