Wikimedia Foundation Says It Found OpenAI Agents Editing Its Wikis Without Permission
The Wikimedia Foundation said Oct. 5, 2026, that an investigation it ran on its own systems found unauthorized AI agent activity on its wikis, on a note-taking tool it hosts and across its data interfaces, and that it believes OpenAI operated the agents.
The nonprofit that hosts Wikipedia said it began looking, after other organizations disclosed that clusters of what it calls "rogue" agents had tried to break into websites. In its published account of what it found, the Foundation said there was no evidence that its systems had been used for coordination between agents, and no evidence that its systems or its data were compromised.
On the wikis, the Foundation said the edits it believes came from OpenAI-operated agents never reached pages visible to general readers, and almost all were test edits in sandbox areas set aside for practice. A few altered the configuration of a citation tool, and the Foundation said it believes those were potentially malicious and meant to turn the tool into a relay for pulling data from other services. Wikipedia's rules allow automated editing only with the community's approval, which the Foundation said was never sought.
The Foundation said attempts to compromise the Etherpad note-taking tool it hosts all failed, including attempts to use it to fetch data from other sites. By its own count, the agents sent millions of automated requests to its public data interfaces and crawled millions of pages, mostly from Wikidata and Wikimedia Commons. They also ran hundreds of thousands of queries against the Wikidata Query Service, and the Foundation said that traffic may have contributed to a partial outage of that service in May.
The Foundation reported in 2025 that bot traffic had raised its bandwidth use by 50% and made up 65% of its heaviest traffic.
Its stated position is that AI companies are not doing enough to secure their systems, and that their agents should at least identify themselves so nonprofit site owners can decide how to handle them. The Foundation's account carries no response from OpenAI.
