Analysis of 520K AI Prompts Puts Blame on Users, Not Chatbots

Journalist Henk van Ess analyzed 522,607 real user questions to ChatGPT, Claude, Gemini, Grok and Copilot, concluding that the chatbots aren't the problem—the users are. Faulty assumptions embedded in prompts often lead to wrong answers.

2026-08-20 ~ 2026-08-20 · 3 related posts

• He read 522,607 real prompts sent to ChatGPT, Claude, Gemini and Grok — henkvaness · 2026-08-20
• Analysis of 522k AI prompts: You are the problem — henkvaness · 2026-08-20
• Study of 522K Questions: AI Errors Often Stem from User's Own Biased Prompts — henkvaness · 2026-08-20