Foreign propaganda groups are finding sneaky ways to slip past traditional newsroom defenses. According to OpenAI, commercial operators based in Iran and Russia recently used ChatGPT to create fake journalist personas, write custom pitch emails, and plant around 100 AI-generated articles in real news outlets around the world.
Instead of building low-quality fake news websites that nobody reads, these groups targeted established media channels. By asking ChatGPT to tweak articles so they matched specific submission guidelines, the operators convinced busy editors to publish political commentaries on international relations without ever verifying who actually wrote them.
Tricking newsroom editors with automated pitch emails
The Iranian operation created seven fake journalist personas complete with AI-generated profile photos and supporting social media accounts. Operatives used ChatGPT to polish English drafts, write plausible author biographies, and compose friendly pitch emails to send directly to newsrooms.
In one case, an article about US military conflict ended up in Florida’s Port St. Joe Star. The publisher told The Washington Post he ran the piece because it came from PeaceVoice, a syndication nonprofit he had worked with for years. PeaceVoice has since tightened its submission policies, and OpenAI permanently banned the accounts tied to the operation (via Digital Trends).
Running a fake think tank in Latin America
At the same time, OpenAI shut down a Russian-linked campaign called “Dark Clark” targeting Latin America. The group set up a fake front organization named the Social Research Center, using ChatGPT to manage tasks and co-opt real local staff who had no idea they were working for foreign handlers.
This Russian network spread fabricated audio clips and fake stories across Bolivia and Ecuador, forcing government officials to issue public denials. OpenAI rated both campaigns as unusually successful compared to typical spam operations. The takeaway for newsrooms is pretty clear: relying on familiar syndication channels or polished emails isn’t enough anymore, and verifying the real person behind a submission is becoming essential.
The post Fake AI Journalists Used ChatGPT to Trick Real American Editors into War Propaganda appeared first on Android Headlines.
