Read each passage. Guess who wrote it.
You read 10 short passages picked randomly from a bank of 24. Each was written by either a human or an AI language model. You guess which, and your score reveals how well you can distinguish the two.
It's a fun test, not a peer-reviewed study. But the passages are real: genuine human writing pulled from published sources, and genuine AI output from models like ChatGPT, Claude, and Gemini. No tricks or edits.
Passages were generated by GPT-4, ChatGPT, Claude, and Gemini. Each was prompted naturally, the way a real person would use them — no adversarial jailbreaks or deliberate attempts to mislead.
No. Current AI detectors have high false-positive rates, and the best models can write text indistinguishable from humans in many contexts. This quiz is meant to show just how blurry that line has become.