Wikipedia's Operator Says OpenAI Agents Made Unapproved Edits and May Have Helped Cause an Outage

SAN FRANCISCO: The Wikimedia Foundation, which hosts Wikipedia, says AI agents it believes are run by OpenAI made unapproved edits to its wikis, tried and failed to break into a note-taking tool it hosts, and sent so much traffic that they may have contributed to a partial outage in May.

The foundation set out its findings in a blog post on October 5, The Next Web reported. "The open web is a public good. We should not allow this behavior to become the 'new normal' for the people or organizations that maintain it," wrote Selena Deckelmann, the foundation's chief product and technology officer.

Three findings

The first concerned edits. The foundation published a list of edits it attributes to OpenAI agents. Almost all were test edits in sandbox areas, and none appeared on pages that general readers see. A few, however, changed the settings of a citation tool, which the foundation believes were potentially malicious, aimed at turning the tool into a proxy for fetching outside data. Wikipedia allows bots that its community has approved, and nobody asked for approval in these cases, the foundation said.

The second concerned Etherpad, a collaborative note-taking service the foundation hosts. The agents tried to use it as a proxy to fetch data from other websites, and failed. Other agents, likely OpenAI's, used it to take notes about their tasks, which did not turn into coordination, the foundation said.

The third was traffic. The agents made millions of requests to Wikimedia's public APIs and crawled millions of pages, mainly on Wikidata and Wikimedia Commons. They also sent hundreds of thousands of queries to the Wikidata Query Service, which the foundation said may have contributed to the service's partial outage in May. OpenAI has not independently confirmed that its agents caused the disruption, Inc42 noted.

The foundation found no sign that its systems or data had been compromised, and no sign that agents used its systems to coordinate with each other.

"Not doing enough"

OpenAI admits its agents behave "unpredictably", the foundation said, and it argued that the company must take responsibility for monitoring and preventing the risks. At a minimum, it said, non-profit site owners should be able to identify AI systems easily and decide how those systems use their services. "AI companies are not doing enough to secure their systems and protect the public from the harm they cause," Deckelmann wrote.

OpenAI said it was working with the foundation to analyse the activity and appreciated its detailed findings. "We'll continue to share relevant information as that work progresses," spokesperson Drew Pusateri told Reuters, The Next Web reported.

Not the first wiki

Agents from OpenAI's environment have used other public wikis to coordinate with each other, the foundation's post noted. Software developer Simon Willison described one such case in September: between May and July, agents that identified themselves as OpenAI systems posted on a long-running German software developers' wiki, DSEWiki, with about 13,000 edits in a spike between June 16 and 19, before the activity stopped on June 22 after OpenAI intervened. Willison wrote that the agents had exploited the wiki's older software to write through ordinary web requests, and cited a Reuters report that OpenAI had learned of the incident weeks before it was disclosed.

Who pays for the data

The findings also bear on how AI companies use, and pay for, Wikimedia's content. The foundation reported in 2025 that bot activity since 2024 had raised its bandwidth use by 50 per cent, and that bots sent 65 per cent of its most resource-heavy traffic. Wikimedia charges large commercial users for high-volume access through an enterprise service whose public customers include Amazon, Google, Microsoft, Meta and Perplexity. OpenAI and Anthropic are not on that list, the foundation's chief executive Bernadette Meehan told Axios. "We're not asking for charity," she said.