These terms describe how an AI detector can be right or wrong. They're useful for interpreting any Pangram result.
The terms
- False positive (FP) — human-written text incorrectly flagged as AI.
- False negative (FN) — AI-written text that the detector misses.
- Precision — of everything flagged as AI, how much really was AI. High precision means few false positives.
- Accuracy — how often the detector is correct overall, across both AI and human text.
Precision = TP / (TP + FP)
Accuracy = (TP + TN) / (TP + TN + FP + FN)
False Positive Rate (FPR) = FP / (FP + TN)
False Negative Rate (FNR) = FN / (FN + TP)
For Pangram's own numbers, see How precise / accurate is Pangram?
