The Wikimedia Foundation confirmed that rogue OpenAI bots made millions of automated requests, potentially causing a partial site outage in May. The bots also attempted unauthorized edits and data scraping. OpenAI spokesperson Drew Pusateri said, "We’re working with them as we review and analyse the activity they identified along with."
The Wikimedia Foundation, the non-profit that hosts Wikipedia, confirmed unauthorised activity by “rogue” OpenAI agents on its platforms, including millions of automated data requests that may have contributed to a partial site outage earlier this year, according to a report by The Verge.
Following recent industry-wide disclosures about AI agents quietly accessing third-party platforms,
Wikimedia revealed that OpenAI bots attempted to exploit its tools and edit wikis, the publication reported. However, the foundation clarified that no systems or data were compromised, nor was there evidence that the bots used Wikimedia infrastructure to coordinate with one another.
Here is what the Wikimedia Foundation discovered:
Unauthorised editing: According to The Verge, Wikimedia identified several edits believed to have been made by OpenAI agents. While most were harmless test edits confined to "sandbox" areas and hidden from general readers, a few targeted the configuration of a citation tool.
The foundation described these as "potentially malicious edits" intended to misuse the tool as a proxy for fetching external data. The bots bypassed Wikipedia policies, which require community disclosure and approval for all automated editing.
Etherpad probing: The AI agents attempted to compromise Wikimedia’s public Etherpad, a community note-taking tool, but were unsuccessful, according to the report.
The bots, it said, attempted to use the tool as a proxy to scrape data from other websites. While some agents took notes on their tasks, this activity did not evolve into coordinated bot behaviour.
Excessive data harvesting: OpenAI agents executed millions of automated requests via public APIs, crawling millions of pages primarily across Wikidata and Wikimedia Commons, The Verge reported.
The bots also made hundreds of thousands of queries to the Wikidata Query Service (WQDS). According to Wikimedia, this unusually heavy traffic may have contributed to a partial outage on the WQDS platform in May.
“The open web is a public good,” the Wikimedia Foundation said in a statement. “We should not allow this behaviour to become the 'new normal' for the people or organisations that maintain it.”
OpenAI responds
OpenAI acknowledged the report and is currently investigating the issue, The Verge reported.
"We appreciate the detailed findings Wikimedia shared with us," OpenAI spokesperson Drew Pusateri said. "We’re working with them as we review and analyse the activity they identified along with our overall investigation, and we’ll continue to share relevant information as that work progresses."
Pusateri added that OpenAI’s internal review has not yet verified whether its agents were directly responsible for the May outage.
