AI writing detectors are terrible. False positives. False negatives. Most are pulling a bait & switch for 'humanization' services.
But spotting AI-written copy? That’s getting easier.
After 2+ years working with LLMs, I’ve noticed five “tells” that show up again and again.
