How To Detect Synthetic Images In Ads

Advertisers use AI images widely, for backgrounds, models, products and whole scenes. Much of it is ordinary creative work. It becomes a problem when an ad makes a product look better than it is, invents results, or fakes a real person's endorsement.

Clues in ad images

• Text on products or signs that's slightly wrong.
• Skin and hair so smooth they look painted, especially in beauty ads.
• Products with impossible features or inconsistent details between ads.
• "Before and after" photos with different lighting, angles or even faces.
• Celebrities endorsing products they've never mentioned elsewhere.

The biggest scam: fake endorsements

Scam ads often use AI-generated or altered images and video of celebrities and public figures to sell investments, supplements and products. Check the person's official accounts: real endorsements are usually mentioned there. If the ad leads to a site imitating a news outlet, it's almost certainly a scam.

Disclosure rules

• Political ads: Google and Meta require advertisers to disclose realistic digitally created or altered content in election and social issue ads.
• Misleading claims: in the US, the FTC treats ads that misrepresent products as deceptive, whatever the method. The UK's Advertising Standards Authority has ruled that beauty filters exaggerating a product's effect are misleading.
• The EU's AI Act adds labeling duties for deepfakes from August 2026.

Checking an ad image

On an advertiser's website, AI Image Detector (made by Patrick Bushe, who runs this site) reads image metadata in Chrome and flags AI tools or tags if present. Ad platforms usually process images, removing metadata, so an empty result is inconclusive. Reverse image search can show whether the same image was used elsewhere, such as a stock or AI art site.

Where to report

• Use "Report ad" on the platform.
• In the US, report scams at ReportFraud.ftc.gov.
• In the UK, complain to the ASA.

See real vs AI images in marketing.