Several chatbots, including ChatGPT, Gemini, DeepSeek and Grok, have been found to cite Russian state propaganda from sanctioned entities, new ISD research finds. 

Reporting on these findings, WIRED draws extensively on ISD’s latest study examining the four widely used chatbots and their responses to questions about the Russian invasion of Ukraine.  

The article highlights key results, including that Russian state-attributed content appeared in 18 percent of responses across all prompts, languages and large language models (LLMs). Certain topics, such as peace talks between Russia and Ukraine, were more likely to trigger references to state-attributed sources.

WIRED also spoke to ISD’s Pablo Maristany de las Casas, who outlined how major tech companies could respond to these findings. 

“It could be providing users with further context, making sure that users understand the times that these domains have a conflict and even understanding why they’re sanctioned in the EU,”Pablo explained. “It’s not only an issue of removal, it’s an issue of contextualising further to help the user understand the sources they’re consuming, especially if these sources are appearing amongst trusted, verified sources.” 

The full article is available on WIRED.