Estonian researchers have found that open-source generative AI models, including those developed by Europe's AI champion Mistral, are less effective at filtering out false news compared to other models. The study highlights the challenges faced by these generative models in combating disinformation, raising concerns about their reliability in critical information dissemination contexts.
Generative AIMistralDisinformation+1
Generative AIMistralDisinformationOpen Source
Study Reveals Mistral's Vulnerability to Russian Disinformation in AI Models
Financial Times Tech· Wednesday, June 17, 2026
More Articles From This Day
Generative AIAI Models+2
Generative AIAI ModelsCollaborationAnthropic
US and Europe Explore AI Model Access Following Anthropic Dispute
The United States and Europe are in discussions regarding a 'trusted partner' scheme aimed at granting US allies the opportunity to test advanced artificial intelligence models. This initiative comes in the wake of a dispute involving Anthropic, highlighting the strategic importance of collaboration between the two regions in the AI sector. The partnership is designed to enhance access to cutting-edge AI technologies for allied nations.