Window ending Sept. 21, 2026

This is a backcast from the current catalogue. It was not a reading published at the time.

Records behind this reading

9 selected records; 2 documented external harms counting (2 qualifying). Records and ratings below reflect the current catalogue.

2 qualifying harm records. Extent undisclosed for both.

Documented harm

2

No harm found (stated scope)

1

No harm reported

5

Impact unknown

0

Alleged, AI role uncorroborated

0

Counts toward the index

2

Unverified, watching

0

Alleged in court

0

Control failure, tracked

6

Tracked separately

1

Documented exclusions: 0 internal; 0 awaiting evidence review. 0 qualifying records with a bounded single-source review.

Inspect the evidence · 8 records

• PI-0087 · Claude Fable 5.1 and Fable 5 used a flaw in a testing partner's sandbox to read files outside itNo harm found (stated scope) · excluded from the harm reading
• PI-0079 · In a researcher demo, planted instructions hijacked Copilot in SQL Server Management Studio to grant sysadmin rightsNo harm reported · excluded from the harm reading
• PI-0074 · Early Claude Opus 4.6 broke into a real outside machine during a hacking test and read one person's dataDocumented harm · qualifying harm · extent undisclosed
• PI-0061 · OpenAI agents put task files on the public internet against instructionsNo harm reported · excluded from the harm reading
• PI-0062 · OpenAI model used a leaked third-party API key, then fabricated the data it could not fetchNo harm reported · excluded from the harm reading
• PI-0063 · OpenAI models wrote notes telling future copies to hide mistakes and ignore constraintsNo harm reported · excluded from the harm reading
• PI-0078 · OpenAI training agents used an internal package repository as a message board across separate samplesNo harm reported · excluded from the harm reading
• PI-0073 · Gemini guessed a password and used leaked credentials to get into three real companies during a hacking testDocumented harm · qualifying harm · extent undisclosed

Status totals cover counted records; aliases, superseded aggregates and records tracked separately are excluded. Eligibility and extent notes overlap those totals. These observations are not statistical uncertainty bounds.

1. PI-0087
   Claude Fable 5.1 and Fable 5 used a flaw in a testing partner's sandbox to read files outside itSept. 1, 2026 · No harm found · Controlled test
2. PI-0079
   In a researcher demo, planted instructions hijacked Copilot in SQL Server Management Studio to grant sysadmin rightsSept. 8, 2026 · No harm reported · Controlled test
3. PI-0080
   Lawyer held in contempt after filing a ChatGPT brief with invented witnesses in a New Mexico murder appealSept. 8, 2026 · Negligible harm · Deployment
4. PI-0074
   Early Claude Opus 4.6 broke into a real outside machine during a hacking test and read one person's dataSept. 9, 2026 · Negligible harm · Controlled test
5. PI-0061
   OpenAI agents put task files on the public internet against instructionsSept. 16, 2026 · No harm reported · Internal research
6. PI-0062
   OpenAI model used a leaked third-party API key, then fabricated the data it could not fetchSept. 16, 2026 · No harm reported · Internal research
7. PI-0063
   OpenAI models wrote notes telling future copies to hide mistakes and ignore constraintsSept. 16, 2026 · No harm reported · Internal research
8. PI-0078
   OpenAI training agents used an internal package repository as a message board across separate samplesSept. 16, 2026 · No harm reported · Internal research
9. PI-0073
   Gemini guessed a password and used leaked credentials to get into three real companies during a hacking testSept. 18, 2026 · Negligible harm · Controlled test

Sources

These links support the records above. Source availability and conclusions may change.

1. Anthropic: System Card: Claude Fable 5.1 & Claude Mythos 5.1Cited in PI-0087
2. Microsoft Security Response Center: CVE-2026-65669: SQL Server Elevation of Privilege VulnerabilityCited in PI-0079
3. Embrace The Red: From SELECT to SYSADMIN with SQL Copilot (CVE-2026-65669)Cited in PI-0079
4. The Hacker News: Microsoft patches record 974 flawsCited in PI-0079
5. New Mexico Supreme Court: Order finding Stephen D. Aarons in direct contempt of courtCited in PI-0080
6. Reuters: ChatGPT invented fake police testimony in murder appeal, New Mexico high court saysCited in PI-0080
7. 404 Media: Lawyer Cites ChatGPT-Invented Fake Witnesses in Murder AppealCited in PI-0080
8. Anthropic: An alignment assessment of recent cybersecurity incidentsCited in PI-0074
9. OpenAI: Our framework for reporting model misalignmentCited in PI-0061, PI-0062, PI-0063
10. OpenAI Alignment: Uploading files to the internet in order to cite themCited in PI-0061
11. OpenAI Alignment: Unauthorized communication via temporary file hosting servicesCited in PI-0061
12. The Hacker News: OpenAI Reveals Six Model Incidents Involving Hidden Failures and Unauthorized UploadsCited in PI-0061
13. NBC News: OpenAI flags 6 new incidents of 'concerning' behavior and unveils plan to track itCited in PI-0061
14. OpenAI Alignment: Signing up for disposable emails and searching GitHub for leaked API keysCited in PI-0062
15. SecurityWeek: OpenAI Says Its Models Searched GitHub for Leaked API Keys During TrainingCited in PI-0062
16. Dark Reading: Rogue Behavior: OpenAI Reveals More Model Misalignment IncidentsCited in PI-0062
17. OpenAI Alignment: Self-generated prompt injections in compaction summariesCited in PI-0063
18. OpenAI Alignment: Encouraging deception in compaction summariesCited in PI-0063
19. Axios: OpenAI discloses six new AI misalignment incidentsCited in PI-0063
20. BleepingComputer: OpenAI details more cases of AI agents taking unauthorized actionsCited in PI-0063
21. OpenAI Alignment: Unsanctioned Artifactory writes and cross-sample communicationCited in PI-0078
22. CSO Online: OpenAI admits six new misalignment incidents under new reporting frameworkCited in PI-0078
23. Irregular: Addressing Recent Incidents: Ongoing Findings and Path ForwardCited in PI-0073
24. TechRadar: Google's Gemini hacked three companies during Irregular AI 'capture-the-flag' testingCited in PI-0073
25. Cybernews: Google's Gemini hacked three real companies during security testCited in PI-0073
26. Al Jazeera (Reuters): Google's Gemini AI hacks 3 companies in security test, then stopsCited in PI-0073
27. GV Wire (Reuters): Gemini Hacked Three Companies in First Known Breakout By Google's AI, WSJ reportsCited in PI-0073

All weekly readings · How this reading is calculated · Download the weekly card