AI & Machine LearningAI Detection

AI Detectors Get It Wrong

Tools claiming to spot AI writing often flag humans

You wrote every word yourself, and a tool says a robot did. What is it actually looking at when it accuses you?

You'll find out

  • What these tools really measure
  • Who gets wrongly accused the most
  • What a score can and can't prove

Why it matters

These scores land on real students and real jobs, so it's worth knowing how little they can prove.

About 3 min6 interactive slides · 5 questionsLesson 543 · Zone 61

Play this lesson

Free to play in your browser or on iPhone

More in this zone

  1. 541Game TheoryParrondo's ParadoxTwo losing games can win when you alternate them
  2. 542AcousticsThe Whispering GalleryWhisper against a curved wall and be heard across a cathedral
  3. 544Time PerceptionTime Slows Down in a Crisis, or Does It?Falling people don't actually see more frames
  4. 545Planetary ScienceWhy Mars Is RedThe whole planet is rusty
  5. 546Computing HistoryAda Lovelace's ProgramThe first published program was written in 1843 for a machine never built
  6. 547ComputabilityTuring's Imaginary MachineA tape, a head and a few rules can compute anything
  7. 548Fractal GeometryThe Dragon CurveFold paper repeatedly and a dragon appears
  8. 549Visual SearchReverse Image SearchA picture can become its own search query