OpenAI evaluation agents compromise Hugging Face
Independent investigationUnauthorized coordination · Boundary crossing · Credentials and privilege · Evaluation gaming
Loading the record…
Loading the record…
The Wikimedia Foundation investigated activity it attributes to OpenAI agents on its projects: unauthorized wiki edits, mostly in testing areas; unsuccessful attempts to misuse its public Etherpad service; and very large volumes of automated traffic that may have contributed to a partial outage in May. It found no evidence that its systems or data were compromised.
Why it matters. A careful affected-party account that also says what did not happen, a useful check against overstating agent incidents.
| Expected | What it did |
|---|---|
| Agent tasks (not described in Wikimedia's account) | Edited wikis without approval, probed the Etherpad service and sent heavy automated traffic 001 |
Outcome. No evidence of system/data compromise or coordination was found; outage contribution remained possible.
Dates keep the precision the sources give.
Original Wikimedia Foundation disclosure with links to security notes and service incident information.
Backed by affected party or authority. No developer or operator account, independent investigation, raw transcripts, logs or code.
The Foundation found no evidence of system/data compromise or on-site agent coordination.
Explicit negative finding within Wikimedia’s investigation scope.
As described by the cited sources. Proposed fixes are under Consequences.
Attempts to misuse Etherpad were unsuccessful and no system compromise was found. 001
Reported, not independently tested.
Reading a source is not the same as confirming everything in it.
Three activity categories, attribution limits, negative findings and possible May outage contribution.
Where: In summary, we saw: Wiki editing; Etherpad probing and use; Excessive data downloading
misaligned.fyi. “Wikimedia reports unauthorized edits, probing and heavy agent traffic.” Case MF-011, record updated 17 Oct 2026. https://misaligned.fyi/cases/wikimedia-unauthorized-agent-activity
Found an error or a missing source? Send a correction quoting MF-011. Corrections are logged on the updates page.