Study finds Mistral more vulnerable to Russian disinformation

Financial Times reported on a study by Estonian researchers finding that Europe's AI champion Mistral was vulnerable to Russian disinformation, with open-source generative models performing worse than other systems at removing false news. The story is a useful research and policy signal because it links model openness, content moderation, and national-security information operations. It also gives the feed a European AI governance angle beyond the dominant U.S. frontier-lab stories.

Key details: Published June 16, 2026, The reported study focused on Mistral and Russian disinformation, Open-source generative models were said to be worse at removing false news than other systems, The finding adds a European model-governance risk to the current feed.

Continue swiping for more AI Brief stories.