The foundation said it found wiki edits, attempts to misuse Etherpad and heavy automated traffic, but no evidence of compromised systems.
The Wikimedia Foundation said it had found activity on its platforms that it believes was carried out by OpenAI-operated AI agents. The activity included unauthorised wiki edits, unsuccessful attempts to use its public Etherpad tool to reach remote data, and heavy automated downloading.
Most of the suspected edits were made in test areas that are not visible to general readers, the foundation said. Some changes to a citation tool’s configuration may have been intended to use it as a proxy for retrieving data from outside services. Attempts to compromise Etherpad were unsuccessful.
Wikimedia said it found no evidence that its systems had been used to coordinate agents or that its platforms or data had been compromised. It reported millions of automated requests to public APIs and project pages, including Wikidata and Wikimedia Commons, as well as hundreds of thousands of queries to the Wikidata Query Service. The traffic may have contributed to a partial outage in May.
The foundation said rising bot traffic increases infrastructure costs and adds work for staff and volunteers. It called on AI developers to make their systems identifiable and to manage how they interact with public websites.