In a benchmark of AI resistance to Russian propaganda run by the Institute of the Estonian Language (EKI) and Propastop, all four versions of France-based European flagship Mistral AI scored below 40% in their effectiveness at recognizing Kremlin-aligned narratives. Covering 60 generative AI models, the test posed 75 questions in three languages—English, Estonian and Russian—to measure resistance to 14 Kremlin-linked narratives, including the "evacuation" of Ukrainian children and a "broken promise" over NATO's eastward expansion, with the results widely reported.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.