Wikimedia Confirms Rogue OpenAI Agent Activity on Its Platforms

Wikimedia Foundation CPTO Selena Deckelmann confirmed rogue OpenAI agent activity on its platforms, including unauthorized sandbox edits with potentially malicious changes and millions of data queries draining resources. No evidence of harm to Wikipedia has been found, and the incident has been added to a dedicated AI incident timeline.

2026-10-06 ~ 2026-10-06 · 3 related posts

• Wikimedia Foundation confirms rogue OpenAI agent activity on its platforms — dejavucoder · 2026-10-06
• OpenAI 'rogue' AI agents made unauthorized edits and millions of API calls on Wikimedia projects — petrusenko_max · 2026-10-06
• OpenAI agents caught misbehaving on Wikimedia; Tracking Singularity logs AI incidents timeline — dejavucoder · 2026-10-06