Operation “Stop News”: Russia-origin influence activity

OpenAI's latest takedown report has a new name in it: Stop News. The company says it banned a cluster of accounts tied to a Russia-origin influence operation that used its tools to mass-produce articles and social posts in several languages, aimed squarely at shaping opinion around the war in Ukraine and broader Western audiences.

What makes this case notable isn't the goal — Russian influence campaigns targeting Ukraine and Europe are old news at this point — it's the workflow. Instead of hiring writers in multiple languages or running clunky translation software, the operators leaned on generative AI to draft content quickly and in bulk, then pushed it out under the Stop News branding across the platforms they controlled. That's the same pattern OpenAI has flagged in prior reports: state-linked actors treating large language models as a force multiplier for propaganda shops that used to need bigger staffs.

OpenAI didn't publish a deep breakdown of reach or engagement numbers this time, but the framing is consistent with its ongoing threat-intelligence work — find accounts abusing its models for coordinated inauthentic behavior, cut off their access, and publish enough detail for researchers and other platforms to spot copies. The company has done this repeatedly over the past two years, banning networks linked to Russia, China, Iran and Israel, among others, all trying to use AI-written text to look more credible or more native to their target language.

And that's really the throughline here. Influence operations haven't changed their objectives — they still want to nudge public sentiment on Ukraine, on NATO, on Western unity. What's changed is the cost of production. A network that once needed a room of people to write convincing Ukrainian, German or English copy can now get a first draft from a chatbot in seconds. OpenAI banning Stop News closes one door, but it's a reminder that the barrier to running this kind of operation keeps dropping, not rising.