An analysis by NPR and NewsGuard found major AI chatbots are largely ineffective at countering foreign propaganda.
OpenAI’s ChatGPT, Google’s Gemini, Microsoft’s Copilot, Meta AI, xAI’s Grok, and Anthropic’s Claude correctly debunked false narratives from Russia, China, and Iran only 75% of the time.
The testing period ran from December 2025 to July 2026.
AI-powered search summaries from Google and Microsoft performed worse than standalone chatbots.
Google’s AI Overviews successfully debunked a majority of false claims.
Microsoft Bing’s AI summaries failed to debunk most of the tested narratives.
The results highlight a critical vulnerability in generative AI models against disinformation campaigns.