AI Risk Atlas Prototype/DemoUnofficial independent experiment. Not an official xAI product. Scores can be wrong.

When the file grew

Signal injection timeline

How public signals entered this desk over time. Seeded cards use the original post date. Live sweep cards use the moment they were collected. Counts are experimental and incomplete.

Methodology — experimental estimates

Scores are automated, experimental estimates from public X posts and a hand-written seed corpus. They are not formal risk assessments, not certified, and not suitable for compliance or operational decisions.

Consequence, likelihood, and urgency are 1–5 judgements applied by this project, not by a standards body. Residual scores assume only the mitigations marked in place. A signed-in reviewer can override residual and mark an item reviewed — that override is still unofficial. Aspect tags (capability, domain knowledge, affordance, impact domain) are a lightweight PRA aid, not a formal hazard analysis.

Full about and disclaimer

All charts →

On file
75
Last 30 days
47
Peak day
08-17
13 cards
Critical share
24%

Daily · last 60 days

Injections by day

A quiet day is not a safe day — it is often just a gap in the sweep.

Severity seepage

What kind of signal arrived

Source seepage

Labs, institutes, press, other

Weekly

Injection rate

Cumulative

File size over time