I Tested a Popular A.I. Slop Detector. It Felt Empowering.
TL;DR
A test of the AI detector Pangram finds the tool convincingly separates chatbot-generated phrasing from human writing. On artificially generated images it performs considerably less reliably, according to the New York Times report. For newsrooms, education and companies that means Pangram can give a useful signal on text checks, but it does not replace provenance records or manual review for images.
Nauti's Take
Pangram is worth a small pilot on your own documents, using known AI-generated and human-written samples to measure false positives. Treat its image detection as a weak signal for now, since the available evidence comes from a single reported test and does not establish reliable provenance.
Sources