Back Lemonde.Fr

Russian disinformation on ChatGPT: Fake Telegram channels, fake videos and a real ...

Lemonde.Fr • August 25, 2026

They hid behind a VPN but OpenAI's security teams managed to spot them. Somewhere in Russia, a small group of operators used ChatGPT accounts to generate propaganda messages in several languages. Most were in English, and the operators regularly asked ChatGPT to disguise linguistic clues that could have revealed they were Russian, the company's security researchers wrote in a detailed report released on Tuesday, August 25. OpenAI's teams found the same messages in various Telegram channels, in English, German, or French, describing a decadent West and portraying Russia as a grand alternative.

But what is even more intriguing is that these user accounts were not only used to churn out the usual clichés of Russian propaganda. They were also heavily used to create text and visuals promoting a think tank, the "International Burke Institute" (IBI), which claims to be based in Israel.

A quick look at IBI's website might suggest it is an internationally recognized institution. It displays a list of prestigious partners, including the UN and UNICEF, and an endless roster of experts supposedly affiliated with the institute, such as former US Secretary of State Mike Pompeo and scholar Francis Fukuyama.

The full story

This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.

OpenAI Disrupts Russian Influence Operation Using ChatGPT OpenAI has banned multiple ChatGPT accounts linked to a covert influence operation originating from Russia. These accounts were used to promote the International Burke Institute (IBI), a fake Israeli think tank, and disseminate pro-Russia narratives across social media platforms. The operation did not rely on malware but instead leveraged generative AI to create posts that appeared academically legitimate. OpenAI's action is part of a broader effort to combat misinformation and influence campaigns on its platform. The company identified these accounts as likely part of a coordinated effort to manipulate public opinion. The operation's scope included various social media platforms, but specific metrics on the reach or engagement of the posts were not disclosed. OpenAI's measures aim to maintain the integrity of its AI systems and prevent misuse for propaganda purposes. 6 sources 45

Extracted Entities

Russian disinformation on ChatGPT: Fake Telegram channels, fake videos and a real ...

The full story

Extracted Entities

Countries (2)

Platforms (2)

Tools (1)