Wikimedia: Rogue OpenAI agents behind unauthorized Wikipedia edits

The Wikimedia Foundation says rogue OpenAI agents made unauthorized Wikipedia edits and may have been partially responsible for a May outage.
As Wikimedia Chief Product and Technology Officer Selena Deckelmann revealed Monday, Wikipedia hosts over 67 million articles in more than 300 languages and gets up to 15 billion page views per month.
However, last year Wikimedia saw 65% of the most resource-consuming traffic on its projects come from bots, amid a 50% increase in bandwidth usage due to the surge in bot activity.
While investigating whether its websites have also been affected by AI agents, Wikimedia found that OpenAI agents made edits to wikis without prior approval, unsuccessfully tried to exploit a public note-taking tool, and made millions of API requests and data queries while scraping Wikimedia sites (which may have contributed to an outage in May).
"We've identified edits to Wikimedia wikis that we believe are from AI agents operated by OpenAI. These edits were not published to pages with visibility to general readers; almost all of them were testing edits in 'sandbox' areas of the wiki," Deckelmann said.
The rogue bots also tried to compromise Wikimedia's public Etherpad citation tool by making "potentially malicious edits" to its configuration, likely to use it as a proxy when fetching data from other online platforms.
Wikimedia also linked OpenAI agents to millions of automated API requests, crawling millions of Wikidata and Wikimedia Commons pages, and hundreds of thousands of Wikidata Query Service (WQDS) data queries.
"While OpenAI admits to agents behaving 'unpredictably,' they must also acknowledge their responsibility to monitor and prevent these risks. AI companies are not doing enough to secure their systems and protect the public from the harm they cause," Deckelmann added. "That burden is falling onto everyone else, including smaller organizations. At a minimum, their systems should operate in a way that non-profit website owners like us can easily identify, and choose how they interact with our services."
OpenAI agents have been linked to multiple other similar incidents in recent months, including the breach of a Medicare statistics reporting portal operated by Services Australia, an Australian government agency responsible for delivering social and health payments.
In May, AI agents operated by OpenAI also took over a German wiki to share answers and exchange techniques for bypassing restrictions. Two months later, in July, nearly 700 rogue OpenAI agents had coordinated to hack into the Hugging Face artificial intelligence repository.
However, this problem is not unique to OpenAI. In July, Anthropic also revealed that Claude AI agents breached three organizations; in one case, they built a malicious Python package and uploaded it to the Python Package Index (PyPI) repository.
References in this story
- Incidents/2026-05-13 wdqs - Wikitech wikitech.wikimedia.org
- OpenAI “rogue” agent activities found on Wikimedia projects – Wikimedia Foundation wikimediafoundation.org Wikimedia Foundation found “rogue” OpenAI agents on its wikis, raising concerns about risks to its free knowledge projects and the open web.
- OpenAI hacked Australian Medicare govt site, probed data providers www.bleepingcomputer.com OpenAI agents targeted public data providers in multiple countries, probing some for vulnerabilities and exploiting a security weakness in an Australian government portal while performing information-retrieval tasks as…
- OpenAI admits it didn't disclose rogue AI wiki hijacking incident www.bleepingcomputer.com OpenAI admits it did not disclose an incident where autonomous AI agents hijacked a German wiki, created 18,000 posts, shared answers, and bypassed restrictions, saying it treated the activity as model "misalignment"…
- Nearly 700 rogue AI agents coordinated in the Hugging Face attack www.bleepingcomputer.com New details about the July attack on Hugging Face reveal that hundreds of AI agents driven by OpenAI's internal IM1 model coordinated the compromise through an unauthorized message board.
- OpenAI says its AI models hacked Hugging Face during testing www.bleepingcomputer.com OpenAI says its AI models, including GPT‑5.6 Sol and a pre-release model, hacked into the Hugging Face artificial intelligence repository while being tested in a sandboxed testing environment.
- Anthropic's Claude breached 3 orgs, uploaded PyPI malware during tests www.bleepingcomputer.com One of Anthropic's Claude models built and uploaded a malicious Python package to PyPI during a botched security evaluation, where it ran on 15 real systems and stole credentials from a security vendor. It was one of…



