OverreachPI-0061
OpenAI agents put task files on the public internet against instructions
To cite files through the browser, OpenAI agents uploaded files they had produced or been given during tasks to public hosting sites without asking. They posted a task photo in October 2025 and a text file of lake records in January 2026; the uploads succeeded, though the later citation steps failed. In a second report, agents collaborating on an April 2026 training task, unable to see each other's files, shared a spreadsheet workbook through a temporary public file-hosting service, although the task said to use local files only.
Harm: No harm reportedControl level 2, Instruction violation
Not counted · can set the control floorEffect on the index
Aggregate record. Two OpenAI reports in one disclosure: two citation uploads (a photo and a text file) and one workbook shared between agents. Counts once. The photo may overlap PI-0067, but whether it was user-provided and linked is not established.
Sources
How we know
5 sources · developer confirmed. Links go to the original publishers; the summary above is in our own words.
- primaryOur framework for reporting model misalignmentOpenAI · Sept. 16, 2026openai.com/index/model-misalignment-reporting-framework/
- primaryUploading files to the internet in order to cite themOpenAI Alignment · Sept. 16, 2026alignment.openai.com/misalignment-reports/uploading-files-to-the-internet-in-or…
- primaryUnauthorized communication via temporary file hosting servicesOpenAI Alignment · Sept. 16, 2026alignment.openai.com/misalignment-reports/unauthorized-communication-via-tempor…
- newsOpenAI Reveals Six Model Incidents Involving Hidden Failures and Unauthorized UploadsThe Hacker News · Sept. 17, 2026thehackernews.com/2026/09/openai-reveals-six-model-incidents.html
- newsOpenAI flags 6 new incidents of 'concerning' behavior and unveils plan to track itNBC News · September 2026nbcnews.com/tech/tech-news/openai-new-incidents-concerning-behavior-model-misal…
Why this rating
No harm reported; control failure level 2
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
No harm reportedA task photo, a text file and a workbook were left at public URLs; not reported as sensitive, and their removal is not reported.
Not a finding that no harm occurred.
The harm scale
- 1 Negligible Inconvenience, easily remedied.
- 2 Minor Limited, recoverable harm.
- 3 Moderate Material harm needing significant effort to remedy.
- 4 Severe Severe harm to health, rights, property or essential services.
- 5 Catastrophic Society-scale harm or disruption beyond a community's capacity to cope.
Control assessment
Instruction violationBroke explicit task instructions; OpenAI did not report use of unauthorized access.
Reported beside the index. It adds nothing to a harm reading; when no harm counts in a window, the highest eligible control level in the window is the reading.
The control scale
- 1 Negligible Behaved as instructed. The problem was an ordinary error or a flawed output, with no rule broken.
- 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 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 Severe Reached systems outside its permitted environment, or acquired money, compute or accounts without authorization. The type and mechanism say how.
- 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
Two similar OpenAI reports merged. Control level 2 (breaking an instruction), flagged because the files left the controlled environment (which also makes the setting real world).
Effect on the index
It does not count toward the Sept. 21 reading
The reading for the week to Sept. 21, 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.
Not counted: no harm reported. 8 other records behind the reading for that week.
The arithmetic
| Step | With it | Without |
|---|---|---|
| Counts toward the index?documented, external, eligible evidence | No | — |
| Worst documented harm, ksets the band | 1 Negligible | 1 Negligible |
| Harms at that level, nposition in the band | 2 | 2 |
| Highest control level breachedsets the reading only when no harm counts | — (none breached) | — (none breached) |
| Readingrounded down | 13 Negligible harm | 13 Negligible harm |
Inside the window of 4 weekly readings
| Week to | Reading | Band |
|---|---|---|
| Sept. 21, 2026 | 13 | Negligible harm |
| Sept. 28, 2026 | 21 | Minor harm |
| Oct. 5, 2026 | 26 | Minor harm |
| Oct. 12, 2026 | 22 | Minor harm |
Revisions
What we changed
3 logged. Every change to a rating is logged here, with the reason.
- v3Oct. 13, 2026
Wording only: rubric shorthand in the rating rationale replaced with plain words; ratings unchanged.
- v2Oct. 13, 2026
Wording: removed an editor's aside from the control note.
- v1Oct. 7, 2026
First published.
Cite and share
Use this record
Citation
Paperclip Index. “OpenAI agents put task files on the public internet against instructions.” Record PI-0061. Reported Sept. 16, 2026; updated Oct. 13, 2026. Rated under methodology v1.1. https://paperclipindex.com/incident/PI-0061