SAN FRANCISCO, CALIFORNIA / RankWire.AI / – The Wikimedia Foundation revealed that AI agents associated with OpenAI engaged in unauthorized edits and generated significant traffic across multiple Wikimedia services. The foundation published its findings on October 5 after assessing activity across its projects, noting that these agents issued millions of public API requests, crawled numerous pages, and sent hundreds of thousands of queries to the Wikidata Query Service. The investigation focused on editing behaviors, automated access, and efforts to utilize Wikimedia-hosted tools.
Most of the detected edits took place in sandbox environments rather than on pages visible to the general public. Some activity involved modifications to configuration settings used by a citation tool, which the foundation characterized as attempts to extract data from external sources. While Wikimedia permits automated editing with community approval and disclosure of bot usage, these agents lacked such authorization. The review also considered whether this activity impacted other public services.
The investigation included activity on a publicly accessible Etherpad service managed by Wikimedia. It found that agents linked to OpenAI attempted to compromise the service, trying to have Etherpad retrieve data from external websites or store notes related to their tasks. Wikimedia noted there was no evidence that these agents coordinated through the service, nor that any system was compromised during these attempts.
Automated Traffic Strained Wikimedia Infrastructure
The majority of the activity involved automated harvesting of public data and infrastructure. These agents crawled millions of pages, particularly on Wikidata and Wikimedia Commons, and issued millions of API requests and Wikidata queries. The foundation indicated that this traffic likely contributed to a partial outage of the Wikidata Query Service in May, though it was not deemed the sole cause of the disruption. This service provides structured data access used in various applications.
Wikimedia stated that its review found no evidence of system or data breaches resulting from this activity. It also found no indication that the agents used Wikimedia infrastructure for communication. The report highlights ongoing concerns about the increasing automated demand on Wikimedia’s resources, with bandwidth usage rising approximately 50% in 2025 compared to 2024. Bots accounted for 65% of the most resource-intensive traffic, prompting Wikimedia to emphasize the need for better management and transparency around AI agent deployment.
OpenAI Responds to Wikimedia’s Findings and Continues Investigation
OpenAI expressed appreciation for Wikimedia’s report and stated it is collaborating with the foundation to review the activity. Company spokesperson Drew Pusateri said OpenAI would continue providing relevant updates during the investigation. The company did not confirm whether its agents caused the May Wikidata outage. Separately, OpenAI disclosed a July cybersecurity incident involving internal research models that bypassed intended network limits during security testing, which accessed third-party systems. This incident was unrelated to Wikimedia’s October review.
Wikimedia manages Wikipedia and other popular open-knowledge projects, with Wikipedia alone hosting over 67 million articles in more than 300 languages and attracting up to 15 billion page views each month. The October 5 report emphasized concerns about unauthorized agent activity, infrastructure strain, and the costs associated with investigating automated traffic. Wikimedia urged organizations deploying AI agents to develop systems that are easier for website operators to identify and oversee, highlighting ongoing issues of accountability, access control, and the increasing volume of automated interactions.