Operation “Stop News”: Recidivist influence activity

This case study was originally published in OpenAI’s October 2025⁠(opens in a new window) report.

Actor

We banned a set of accounts that were attempting to use our models to generate content for the covert influence operation we wrote about in October 2024 as “Stop News”. This activity originated in Russia and appeared consistent with an operation run by a marketing company. Our investigation identified efforts by this actor to use other companies’ AI tools to generate videos that were then posted on YouTube and TikTok. We are not able to independently confirm which models they eventually used.

Behavior

Similar to the activity we disrupted in October 2024, this operation mainly used our models to generate content which was then posted on social media and a set of websites that posed as the same news outlets in Africa and the UK that we wrote about last year.

While this activity appeared to be an attempt to continue the earlier operation following our and industry peers’ disruptions, we identified some notable differences. First, the proportion of image generation was significantly lower. Where the original Stop News activity was, as we noted at the time, unusually prolific in its use of imagery, the latest activity focused more on text generation, with image generation being relatively rare. This is consistent with the assessment by the French digital agency, VIGINUM, that “since the publication of the Meta and OpenAI reports, the use of AI-generated images has fallen drastically”.

Second, the operation also used ChatGPT to generate scripts and descriptions for news-style short videos. Typically, the operators would input a lengthy Russian-language text and ask to generate a video script from it. They would then ask the model to translate the text into French. Third, they would ask the model to generate an SEO-optimized description and hashtags.

Completions

Tweet by operation’s main Africa-focused X account. The image was generated using our model.