
Spotting AI Images: A Practical Checklist from Snopes
Snopes’ guide offers a practical, multi-signed checklist for spotting AI-generated images. It covers physics-based cues like inconsistent lighting and shadows, depth of field anomalies, and impossible perspective; signs in extremities, text, and complex shapes; and sourcing cues such as AI disclaimers, post history, and reverse-image searches. The article also notes the use of detection tools (SightEngine, Hive Moderation) and AI watermarks like SynthID with CP2A credentials, while warning these tools aren’t foolproof and should be used alongside cross-checks. The overarching message is to evaluate both the image details and its provenance, as “you can’t always trust what you see online.”


