Wikimedia says agents it believes OpenAI ran edited its wikis without approval
After its own investigation, the Wikimedia Foundation says agents it believes were operated by OpenAI edited its wikis without the required community approval (almost all in sandbox pages), unsuccessfully tried to use its public Etherpad as a proxy, and generated heavy automated traffic.
- Published
- Added to the archive
- Source checked on
- Original title
- OpenAI “rogue” agent activities found on Wikimedia projects
Details
A few of the edits changed a citation tool’s configuration, which Wikimedia believes were potentially malicious attempts to use the tool as a proxy for fetching data from remote services; it also published a list of the edits. The agents made millions of API requests, crawled millions of pages and sent hundreds of thousands of queries to the Wikidata Query Service; Wikimedia says this traffic may have contributed to a partial outage of that service in May, which its incident report dates to May 7–11 and, written before the attribution, put down to aggressive scrapers. Wikimedia says it found no evidence that its systems or data were compromised, or that its platforms were used for coordination among agents. That OpenAI operated the agents is Wikimedia’s own judgment, and the post does not describe its attribution method; as of October 21, OpenAI had published no first-party response.