AI Risk Wire

Cases › operational mistake › LR-2026-0468

Sacramento Bee publishes AI-generated false earthquake and wildfire reports

Preliminary Why: Publisher not yet named · subject to change

Status: action reported, no fix

Severity moderate
Reported 26 Mar 2026
Industry news publishing

No source states: loss or records, developer and model, agent tool, could do, deployer.

The Sacramento Bee publishes AI-generated real-time wildfire and earthquake updates produced by automated bots built on newsroom templates. Over a few days its website published several false earthquake reports, which the USGS's automated notification system had generated and then retracted, plus one false wildfire report about a Ventura County fire that did not exist. The wildfire error appears to have come from data supplied by the National Interagency Fire Center. Cal Fire noticed the erroneous report and notified the newspaper. A correction now appears on the wildfire post, while some earthquake posts remained live. The disclaimer on the posts was later revised to describe AI use and journalist review.

What happened

1 · What it was doingThe newspaper was publishing AI-generated real-time wildfire and earthquake updates from automated bots.
2 · What went wrongSeveral false earthquake reports and one false wildfire report were published.
3 · What it reached or didSome false earthquake posts remained live on the site.
4 · How it was caughtCal Fire noticed the erroneous wildfire report and notified the newspaper.

What was done

A correction was added to the wildfire post and the disclaimer was changed.

Risks

Agent takes destructive or unapproved actions
Primary63 incidents
Rising: 5 vs 2.2 a week

Sorted by rule from the case's fields; not yet read from the sources.

Controls

Not yet read for controls. Controls that mitigate the primary risk:

Harness

Tool allow-list · Fail-closed defaults · Stop control

Application

Human approval for high-impact actions · Change review and separation of duties · Least-privilege access for the agent · Production kept apart from test · Activity logs · Adversarial testing before deployment

Infrastructure / target

Backups and recovery

Similar incidents

Bee Cheng Hiang mass marketing e-mail exposes 95,000 customer addresses after AI-generated code omitted recipient hiding2026-09-30

All cases of this kind ›

Evidence

No source states: when it happened, when it was found, how many records were exposed, the direct loss, a count of the other parties affected, any court case, the cause, what set it off.

Show the full account: 7 items from the primary sources

What happened, in order

Not dated in the sources:

  • False earthquake reports were attributed to the CA Earthquake Bot.
  • The USGS removed the erroneous earthquake notices from its website.
  • The automated error mistook a New Mexico wildfire for a Ventura County fire.
  • A correction appeared on the wildfire post.
  • Cal Fire saw the erroneous report and notified the newspaper.
  • Earthquake posts still linked to USGS pages that had since been deleted.
  • The disclaimer on the posts was changed.
Show every field, with its source and grade

The AI system

What it was for
Publishing automated earthquake and wildfire news articles from structured data on templates created by journalists
C one secondary source states it

What happened

Earliest source we hold
2026-03-26
D inferred by us from the described behaviour, reason shown
What went wrong
operational mistake
D inferred by us from the described behaviour, reason shown

Impact

Severity
moderate
D inferred by us from the described behaviour, reason shown
Can it be undone
partly
C one secondary source states it

Context

Industry
news publishing
C one secondary source states it
Country
United States
C one secondary source states it

Notes on gaps

Counts of posts, readers, and records are not stated.

LR-2026-0468 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON