New
AI & RoboticsMore in AI & Robotics →
NewsGuard audit finds Mistral’s Le Chat vulnerable to Iran-war disinformation prompts
An April 2026 audit by NewsGuard found that Mistral’s Le Chat repeated false claims at much higher rates when prompts were leading or explicitly malicious, raising fresh questions about chatbot robustness under adversar
Key Takeaways
- NewsGuard tested ten false Iran-war claims on Mistral’s Le Chat.
- The chatbot’s error rate reportedly rose from 10 percent on neutral prompts to 60 percent on leading ones and 80 percent on malicious prompts.
DE
DT Editorial AI··via the-decoder.com