Overview
- OpenAI disclosed Thursday that it banned clusters of ChatGPT accounts tied to two coordinated campaigns after tracing activity back to at least July 2025 and documenting a surge in output this year.
- The Russian-linked cluster called Dark Clark ran a purported Social Research Center in Latin America that co-opted unwitting local staff to produce original content and earned a top impact rating of 5 on OpenAI’s IO Breakout scale.
- The Iran-linked operation, dubbed Bogus Bylines, used seven fake Western journalist personas to place almost 100 articles across roughly a dozen small-to-medium outlets and received a Category 4 impact rating.
- OpenAI says operators paired human handlers with ChatGPT to draft long-form pieces, write editorial pitches in Farsi and English, and generate social-media comments, exploiting weak contributor vetting at targeted publications.
- The company published its findings and triggered outlet reviews and platform suspensions, leaving key questions about who contracted the Iranian campaign and how newsrooms should tighten verification going forward.