An investigation by NewsGuard found that 10 leading chatbots generate Russian disinformation narratives from state-affiliated websites.

The authors analysed 570 prompts (57 prompts per chatbot) and concluded that the false Russian disinformation narratives appeared in 152 prompts, 29 responses repeated the false claim with a disclaimer, and 389 responses contained no misinformation — either because the chatbot refused to respond (144) or it provided a debunk (245). 

Business and Human Rights Resource Centre invited OpenAI, You.com, xAI, Inflection, Mistral, Microsoft, Meta, Anthropic, Google and Perplexity to respond. None of the companies did.