How to Spot AI-Edited Images: A Non-Technical Checklist

AI image editing is getting better fast. You don’t need to become a forensic analyst to protect yourself, but you do need a reliable checklist. Below is a pragmatic way to evaluate suspicious images before you share them or trust them.

✅

Goal: reduce false confidence. You are not trying to “prove” an image is fake in 10 seconds—you are deciding whether it is trustworthy enough to act on.

Step 1: Start with context, not pixels

Who posted it first? Anonymous reposts are a red flag.
Why now? Viral timing often correlates with manipulation.
What is the claim? The more explosive the claim, the higher the incentive to fabricate.

Step 2: Look for “cheap” inconsistencies

These are mistakes that still show up even in high-quality synthetic media:

Hands and jewelry: warped fingers, duplicated rings, inconsistent nail shapes.
Text and logos: misspellings, uneven kerning, wrong brand marks.
Background geometry: crooked frames, impossible reflections, melted patterns.
Lighting: highlights that don’t match the light source; shadows going different directions.

Step 3: Check the “story” inside the image

Does clothing match the setting?
Does the posture look physically plausible?
Do objects obey gravity and perspective?

If multiple elements feel off in different ways, that’s often a stronger signal than any single artifact.

Step 4: Do a quick verification loop

Reverse image search (if possible): see if an earlier version exists.
Cross-platform check: is the same source posting elsewhere with consistent details?
Look for original media: longer video, behind-the-scenes photo, or a higher resolution upload.

💡

If you are verifying a potentially harmful image (harassment, intimate imagery, doxxing), avoid “spreading to verify.” Save evidence privately and report via proper channels.

Step 5: Choose a safe action

When you can’t verify quickly:

Don’t repost.
Don’t tag the subject (it can amplify harm).
Ask for a source—politely, and in public if appropriate.

Closing thoughts

The best defense is not perfect detection—it’s good habits: slow down, ask for provenance, and avoid amplifying content that could hurt someone.