1. Run an AI image analysis
Start with an automated detector to identify possible synthetic-image signals. Treat the result as one piece of evidence rather than a final verdict.
No single visual clue or detector can reliably identify every AI-generated image. Combine several independent checks to reach a better-informed assessment.
Upload an image to inspect AI-generation signals, technical metadata and available content credentials.
JPG, PNG, WebP · Maximum 10 MB
JPG, PNG & WebP · Up to 10 MB · No account required
THE EXPLAINER
Understand the available evidence and its limitations.
Start with an automated detector to identify possible synthetic-image signals. Treat the result as one piece of evidence rather than a final verdict.
EXIF information may contain camera details, software names, or timestamps. Metadata can be changed or removed, so its presence or absence is not conclusive.
Some images include C2PA provenance records describing creation or editing steps. A credential must be validated before its claims can be treated as authenticated.
Find the earliest credible publication, compare related images, and check whether the surrounding story is supported by independent sources.
COMMON QUESTIONS
Clear answers about what image verification can and cannot establish.
Sometimes visual inconsistencies raise suspicion, but modern AI images can appear realistic. Visual inspection alone is unreliable.
No. Detectors can produce both false positives and false negatives.
VERIFY WITH CONTEXT
AI detection is an estimate. Metadata and provenance provide additional context, not absolute proof.