OpenAI says it banned Russia- and Iran-origin influence operations that used ChatGPT alongside conventional tactics to conceal who was behind their messages. Its 8 October report identifies real publications as an important distribution route.
Based on 2 primary sources · See sources
Different uses of AI, different routes to readers
The report combines account activity with public-source checks. These are OpenAI’s findings, rather than an independent audit of every placement.
What OpenAI reported
Company findings · 8 October 2026
| Operation | AI’s reported role | Reported distribution |
|---|---|---|
| Russia-origin / Dark Clark | Mainly internal reporting | Think-tank front; some Ukraine-related fakes reached Peruvian and Polish media |
| Iran-origin / Bogus Bylines | Seven false bylines; AI refined articles and pitches | Almost 100 articles published or syndicated; weak engagement on identified social replies |
An audience claim is not an impact measurement
One revealing detail: Iranian operators counted views of posts they replied to as evidence of their own impact. Those views did not establish that anyone saw—or believed—their replies.
The Breakout Scale cited by OpenAI describes how material travels, including amplification by media and public figures. Researcher Ben Nimmo introduced the framework in 2020. It is a distribution framework, rather than a survey showing that people’s opinions changed.
Knowing how content was made is a separate question from knowing who is promoting it. Our look at provenance and truth develops that distinction.
Sources
- False-front operations and observed distribution
OpenAI · Research ·
- The Breakout Scale and its distribution categories
Brookings / Ben Nimmo · Research ·
Written with AI assistance from the sources above. Analysis and practical implications are our interpretation. Our editorial approach.
