OpenAI announced that it had shut down two ChatGPT accounts that were part of foreign influence operations. One campaign, called “Dark Clark,” was run by Russian actors and aimed to spread false stories about Ukraine in Latin America. The other, named “Bogus Bylines,” was an Iranian effort that placed about 100 fabricated articles on the US‑Iran conflict in online news sites around the world.
How the operations worked
Both groups created fake journalist personas and used AI tools to write and translate the articles. The Russian team also fabricated audio clips and documents that prompted fact‑checks and official denials in Ecuador and Peru. The Iranian side added social‑media comments, but those posts received little engagement. OpenAI says the Russian operation reached a “category 5” rating on its Breakout Scale because politicians reacted to the content, marking the first such high‑level case in 2½ years.
OpenAI’s response
OpenAI’s investigation led to the immediate banning of the two ChatGPT accounts used in the campaigns. The company has not disclosed whether any other accounts were involved. It also released a report describing the tactics and the role of AI in the ops, emphasizing that the influence efforts targeted legitimate news outlets rather than just social media platforms.
Why it matters
For regular readers, this shows that AI tools can be misused to insert false stories into trusted news sites, not just social feeds. OpenAI’s account bans may limit the spread of such fabricated content, but the report also highlights that detecting AI‑driven disinformation remains a challenge.
