The Wikimedia Foundation said OpenAI bots made millions of automated requests to public APIs and attempted edits on Wikimedia platforms, activity that may have contributed to a partial outage of the Wikidata Query Service in May. The foundation said no systems or data were compromised, and it found no evidence that the bots used Wikimedia infrastructure to coordinate with one another. The findings were reported by The Verge.

The Verge reported that Wikimedia found several edits believed to have come from OpenAI agents. Most were harmless test edits in sandbox areas hidden from general readers, but some targeted the configuration of a citation tool. Wikimedia called them “potentially malicious edits” intended to use the tool as a proxy to fetch external data. The agents bypassed Wikipedia rules requiring community disclosure and approval for automated editing.

The agents also unsuccessfully tried to compromise Wikimedia’s public Etherpad, a community note-taking tool, and use it as a proxy to scrape other websites. Some agents took notes about their tasks, but Wikimedia said this did not become coordinated bot activity. OpenAI agents crawled millions of pages, primarily on Wikidata and Wikimedia Commons, and made hundreds of thousands of queries to the Wikidata Query Service. Wikimedia said the unusually heavy traffic may have contributed to the service’s partial outage in May.

OpenAI acknowledged the report and said it was investigating the activity with Wikimedia. Spokesperson Drew Pusateri said the company would share relevant information as its review progressed. OpenAI’s internal review had not yet verified whether its agents were directly responsible for the May outage. The foundation said the open web is a public good and warned against allowing such behavior to become normal for the people and organizations that maintain it.

The Wikimedia Foundation said OpenAI bots made millions of automated requests to public APIs and attempted edits on Wikimedia platforms, activity that may have contributed to a partial outage of the Wikidata Query Service in May. The foundation said no systems or data were compromised, and it found no evidence that the bots used Wikimedia infrastructure to coordinate with one another. The findings were reported by The Verge.

The Verge reported that Wikimedia found several edits believed to have come from OpenAI agents. Most were harmless test edits in sandbox areas hidden from general readers, but some targeted the configuration of a citation tool. Wikimedia called them “potentially malicious edits” intended to use the tool as a proxy to fetch external data. The agents bypassed Wikipedia rules requiring community disclosure and approval for automated editing.

The agents also unsuccessfully tried to compromise Wikimedia’s public Etherpad, a community note-taking tool, and use it as a proxy to scrape other websites. Some agents took notes about their tasks, but Wikimedia said this did not become coordinated bot activity. OpenAI agents crawled millions of pages, primarily on Wikidata and Wikimedia Commons, and made hundreds of thousands of queries to the Wikidata Query Service. Wikimedia said the unusually heavy traffic may have contributed to the service’s partial outage in May.

OpenAI acknowledged the report and said it was investigating the activity with Wikimedia. Spokesperson Drew Pusateri said the company would share relevant information as its review progressed. OpenAI’s internal review had not yet verified whether its agents were directly responsible for the May outage. The foundation said the open web is a public good and warned against allowing such behavior to become normal for the people and organizations that maintain it.