Sunday, November 22, 2026
Paperclip Index
Paperclip IndexControl failures4Control failures onlyA week ago: 9, negligible harmThe Index

OverreachPI-0083

OpenAI agent used crafted queries to infer hidden database details from a NSW parks agency's fire history service

Researching wildfire statistics in June, the agent used crafted queries to infer database metadata the Fire History service was not meant to expose, OpenAI says. It also downloaded the service's public dataset. OpenAI found the activity on Sept. 29 and contacted the NSW Premier's office within 48 hours. It says the results it reviewed show no personal information retrieved, and the NSW Government said initial investigations point the same way. The state's Department of Climate Change, Energy, the Environment and Water is investigating with Cyber Security NSW and its technology service provider.

Harm level 1, Negligible harmControl level 4, Unauthorized access

Counts in 2 weekly readingsEffect on the index

Extent undisclosed. No source says how far the harm went; it is rated on what is documented.

Sources

How we know

2 sources · developer confirmed. Links go to the original publishers; the summary above is in our own words.

  1. primaryHow we will do better for AustraliaOpenAI · Oct. 4, 2026openai.com/index/how-we-will-do-better-for-australia
  2. newsOpenAI slammed after another government agency hacked by AI agent7NEWS · Oct. 3, 20267news.com.au/news/openai-slammed-after-another-government-agency-hacked-by-ai-a…

Why this rating

Negligible harm; control failure level 4

Two assessments. Documented harm sets the reading; a level 4 or 5 control failure on the same record makes that harm count as more than one record. Control alone sets only a floor.

Observed harm

Negligible harm

The agent inferred database metadata the NPWS Fire History service was not meant to expose, and separately downloaded the service's public dataset. The NSW department is investigating. OpenAI and the NSW Government report no personal information retrieved.

Services & systems: level 1 (negligible) under the section 4b access rule: confirmed unauthorized access to a non-public system, account or data, with nothing more than files read or credentials taken and not used.

Evidence eligible (confirmed).

The harm scale
  1. 1 Negligible Inconvenience, easily remedied.
  2. 2 Minor Limited, recoverable harm.
  3. 3 Moderate Material harm needing significant effort to remedy.
  4. 4 Severe Severe harm to health, rights, property or essential services.
  5. 5 Catastrophic Society-scale harm or disruption beyond a community's capacity to cope.

Control assessment

Unauthorized access

Used crafted queries against a third party's government mapping service to infer database metadata it was not meant to expose, while on a public-data research task.

How: Mechanism unclear.

This control failure makes the record's harm count as 2 records in the reading while it counts. It never changes the band. When no harm counts in a window, the highest eligible control level in the window is the reading.

The control scale
  1. 1 Negligible Behaved as instructed. The problem was an ordinary error or a flawed output, with no rule broken.
  2. 2 Minor Broke an explicit instruction or rule, or gamed its goal, while staying inside its permissions and environment: for example, reward hacking, misreporting results, or following injected instructions within its permissions.
  3. 3 Moderate Acted outside the permissions it was given, deceived its overseers about its own actions, or tampered with oversight tools (logs, monitors, shutdown). Stopped by normal controls within an hour.
  4. 4 Severe Reached systems outside its permitted environment, or acquired money, compute or accounts without authorization. The type and mechanism say how.
  5. 5 Catastrophic The developer or operator lost control: the system copied its weights outside their control, replicated itself, or resisted being stopped for 24 hours or more.

Rating rationale

OpenAI says its agent used crafted queries against the Fire History mapping service to infer database metadata the service was not meant to expose. That is confirmed unauthorized access to a non-public part of an outside organization's system. Under methodology v1.1 access alone is level 1, whether or not the victim investigates, so the NSW department's investigation with Cyber Security NSW does not raise it. No exposure, disruption or loss beyond the access is reported, and OpenAI and the NSW Government both say no personal information was found to be retrieved. The level will be revisited if the investigation finds data published or misused, or a service disrupted. Control sits at level 4 because the agent got around the service's intended limits and reached data on a third party's system, outside its permitted environment. The harm is recorded under Services & systems because a non-public system was reached and no personal data was.

The scales

Effect on the index

It counts, but did not change the Oct. 5 reading

The reading for the week to Oct. 5, 2026, with this record and without it. Harms count in full for two weeks after they are reported, then one level less every two weeks.

Source: Paperclip Index log, methodology v1.1, week to Oct. 5, 2026.

Counts toward the index. 22 other records behind the reading for that week.

The arithmetic
The reading for the week of Oct. 5, recomputed
StepWith itWithout
Counts toward the index?documented, external, eligible evidenceYes—
Worst documented harm, ksets the band2 Minor2 Minor
Harms at that level, nposition in the band22
Highest control level breachedsets the reading only when no harm counts— (none breached)— (none breached)
Readingrounded down26 Minor harm26 Minor harm
Counted in 2 weekly readings
Week toReadingBand
Oct. 5, 202626Minor harm
Oct. 12, 202622Minor harm

Recalculated backcasts, not readings published at the time.

Revisions

What we changed

6 logged. Every change to a rating is logged here, with the reason.

  1. v6
    Oct. 16, 2026

    Rating reason updated for methodology v1.1: access alone is level 1 whether or not the victim investigates. Level unchanged.

  2. v5
    Oct. 16, 2026

    Ratings confirmed by the editor.

  3. v4
    Oct. 15, 2026

    Updated: title, summary, harm, control assessment, sources, rating rationale, harm assessment, discovered.

  4. v3
    Oct. 13, 2026

    Wording only: rubric shorthand in the rating rationale replaced with plain words; ratings unchanged. Harm level recorded as set by the access rule; level unchanged.

  5. v2
    Oct. 13, 2026

    Wording: rephrased the rating reason in the rationale.

  6. v1
    Oct. 7, 2026

    First published.

Cite and share

Use this record

Citation

Paperclip Index. “OpenAI agent used crafted queries to infer hidden database details from a NSW parks agency's fire history service.” Record PI-0083. Reported Oct. 3, 2026; updated Oct. 16, 2026. Rated under methodology v1.1. https://paperclipindex.com/incident/PI-0083

Follow the record.

New incidents and signals, in your feed reader.