Sakhanda Wire
NVDA $152.94 -2.05% MSFT $505.64 -2.38% GOOGL $280.11 -1.27% META $619.47 -1.11% AMZN $256.54 -0.47%
← Back to the news

AI detectors rely on uncertain judgments about authorship

Tools such as GPTZero, Pangram, and Turnitin’s AI detector do not primarily establish authorship by matching passages against material published online. Instead, they use their own AI models to estimate whether a text may have been produced by a person or by AI. GPTZero says its system examines wording, rhythm, and structure, along with patterns in length and tone that it associates with AI-generated writing.

The assessment is comparatively subjective and less verifiable than finding an exact or similar passage online. The process can also misidentify writing by people who speak English as a second language. Turnitin has said that its detector incorrectly labels less than 1 percent of human-written material as AI-generated. Pangram claims a false-positive rate of 1 in 10,000, while GPTZero says its rate of mistaking human writing for AI is similarly low. Those figures are claims made by the companies, while the underlying method remains an AI-based prediction rather than direct proof of authorship.

Written by Sakhanda Wire AI from reporting by The Verge. Facts, quotes and full context are in the original article below.

Originally published by The Verge on

Read the original on The Verge ↗

The reporting and the image belong to The Verge and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news