We Stress-Tested a Live AI Chatbot. It Invented a Fake Identity, Cited Fake Statistics, and Scored 55/100.
A live AI chatbot built on Groq's API scored 55/100 (Grade E) in a stress test using BotCritic, revealing critical failures including invented identity, fabricated statistics, and inability to handle context. The bot fal…