Highly persistent OpenAI internal model: Unnamed government statistics website
Review CAT-R05 · Catalogue R05, RAT-003
Developer-confirmed intrusion. Ordinary hospital-bed research task; exploit not requested. Agency and country redacted. Do not assume this is Medicare or count as a named country.
Reviewed: Not recorded · Rule: external-intrusion-v1
Same government-statistics exploitation as RW-0031/R05. Primary confirms file disclosure, secrets retrieval and authorization bypass; no new event.
Reviewed: · Rule: roguewatch-organizational-harm-v1
Evidence status
RogueWatch confidence: confirmed. Developer-confirmed intrusion
Dates and discovery
- Incident date
- 2026-06-17
- Catalogue disclosure / source date
- Added to research / record
- Posted to RogueWatch
- RW-0031:
- First observed
- Not recorded
- Discovery channel
- Not recorded
Related incident records: RW-0031 — published incident
Dates use America/Los_Angeles. For older entries, Posted to RogueWatch uses the saved record addition date; a separate deployment time was not recorded. Future posting timestamps record first public qualification. Source publication and incident occurrence are separate.
Sources
- OpenAI Alignment — Retrieving server secrets and bypassing authorization on a government statistics website (Oct 15, 2026)Original account · alignment.openai.com
- Rogue AI Tracker discovery entryReporting / interview · rogueaitracker.com
- OpenAI model broke into a government website during trainingRelationship not established · www.notebookcheck.net · Published 2026-10-18
Decision history
- · Assessment imported
Imported the existing research decision. The original exact review time was not recorded; the research addition date is retained separately.
Evidence source - · Case-by-case comparison audit
RAT-003: existing. Same government-statistics exploitation as RW-0031/R05. Primary confirms file disclosure, secrets retrieval and authorization bypass; no new event. Source checks: read: https://alignment.openai.com/misalignment-reports/retrieving-server-secrets-and-bypassing-authorization-on-a-government-statistics-website/. Same government-statistics exploitation as RW-0031/R05. Primary confirms file disclosure, secrets retrieval and authorization bypass; no new event.
Evidence source - · Incident record added
RW-0031. Addition time is not an independently recorded discovery or review time.
Evidence source · Evidence source - · Incident correction
impact.records_exposed: 1000 → Exact count undisclosed; source says more than 1,000 directory records. The source gives a lower bound, not an exact total. Preserve the 1,000+ figure in the narrative without encoding it as an exact count.
Evidence source
A review decision applies only to the named scope. Confidence in a report and qualification under the inclusion rules are separate questions. Review records do not add incidents to the globe or leaderboard.