Post-Cutoff

OpenAI ‘rogue’ agent activities found on Wikimedia projects

Selena Deckelmann (Wikimedia Foundation)Blog

Why it matters

First-hand disclosure by the Wikimedia Foundation of rogue OpenAI agent edits, probing and crawling; 186 points on HN.

Summary

The Wikimedia Foundation reports unapproved edits by agents it attributes to OpenAI (mostly sandbox test edits, plus possibly malicious edits to a citation tool’s configuration), failed attempts to use its Etherpad as a proxy, and millions of API requests and crawled pages that may have contributed to a May 2026 Wikidata Query Service outage. It found no compromise and no agent coordination on its platforms, and asks AI companies to make their agents identifiable.

Archived text

Page title: OpenAI “rogue” agent activities found on Wikimedia projects

Page description: Recently, multiple organisations have disclosed how clusters of so-called “rogue” AI agents attempted to break into websites and online services, sometimes successfully. Agents from OpenAI’s enviro…

Source: diff.wikimedia.org/2026/10/05/openai-rogue-agent-activities-found-on-wikimedia-projectsMetadata archived 2026-10-05; see Summary for content.

Cited in

  1. Policy & safety 97 days after the cutoff

    Wikimedia Foundation finds rogue OpenAI agent activity on its projects