AI Detector
An AI detector is a machine-learning classifier that estimates the probability a piece of text was generated by a language model.
What detectors look at
- Perplexity, how predictable each next word is.
- Burstiness, variance in sentence length and complexity.
- N-gram fingerprints, common AI phrases like "delve into" or "it is important to note".
- Structural signals, paragraph length symmetry, rule-of-three patterns, transition-word density.
- A trained binary classifier that combines all of the above.
Common detectors
Widely used AI detection tools include GPTZero, Originality.ai, Copyleaks, ZeroGPT, Sapling, and Turnitin's AI writing detection. Each weighs the signals above differently, so the same text often gets different scores across tools.
Important caveats
AI detectors are statistical, not authoritative. They produce false positives on human writing, especially formal, technical, or non-native-English prose. They also miss edited AI output. See AI Detector False Positives for the research.