Overview
- OpenAI disclosed Thursday, Oct. 8 that it had banned clusters of ChatGPT accounts tied to two campaigns and published a disruption report that it shared with authorities.
- The Iran‑linked operation, labeled “Bogus Bylines,” used seven fake journalist personas to place about 100 articles across roughly a dozen small and mid‑size outlets while prompting the model in Persian to draft copy and pitches.
- The Russia‑linked effort, called “Dark Clark,” ran a fake Social Research Center that appears to have co‑opted unwitting local staff in Latin America to produce content aimed at damaging Ukraine’s reputation.
- OpenAI found common tradecraft: operators used VPNs, AI to translate and polish English, AI‑generated author photos and social accounts, and bulk social‑media comments to amplify pieces and exploit light vetting by syndicators and smaller newsrooms.
- The company rated Dark Clark a Category 5 and Bogus Bylines a Category 4 on its influence scale, but it could not definitively identify sponsors for the Iran campaign, a gap that leaves enforcement, sanctions and newsroom safeguards unresolved.