Think you can tell the difference between AI and human writing? Take the quiz and find out. 10 rounds — each one trickier than the last.
You'll see 10 text samples. Guess whether each was written by AI or a human. See your score and learn which patterns gave it away.
Research into human AI-detection consistently finds people perform close to guessing, and that confidence does not track accuracy — those who feel certain are not much likelier to be right. Two errors dominate. We over-flag careful human writing, particularly formal prose and writing by non-native English speakers, because clean structure reads as machine-like. And we under-flag conversational AI text, because most people's mental picture of "AI writing" is the stiff default register rather than what a model produces when asked to write casually.
The reliable signal is not vocabulary or polish — it is specific, verifiable detail drawn from experience. A model writes "many businesses see improved retention"; a person writes "we lost three clients in Q2 before we found the onboarding email was going to spam." Supporting signals include uniform sentence length, hedging applied evenly to every claim, transition words opening most paragraphs, and an unwillingness to commit to any position. None is conclusive alone; they matter in combination.
Want a statistical second opinion instead of a guess? Paste text into the AI Content Detector, or read our guide to the signs that text was written by AI.