Resemble AI finds 8.9% of scanned Meta ad files carry synthetic media, with disclosure at 0.06%

Resemble AI published its 2026 Synthetic Advertising Report on 22 October 2026, based on scanning 37,768 image and video files from 15,433 ads run by 509 companies across 26 industries on Facebook and Instagram. The files were scanned with the company's own detection engine, Resemble Detect.

Company-reported findings:

• 8.9% of files, roughly one in eleven, contained AI-generated or manipulated media.
• 222 of 389 brands ran at least one flagged file. Among brands that used AI more than once, about 12.5% of their ads were synthetic.
• 70% of the flagged ads were photoreal, and 46% depicted a person.
• 2 of 3,379 flagged files (0.06%) mentioned AI anywhere in the ad copy. None of the 971 files checked carried C2PA content credentials.
• In a second scan covering nine sportsbook, prediction-market and crypto brands, 62 of 142 detected fakes (44%) were caught through audio alone.
• Crypto and brokerage was the most synthetic sector, at 29.3%.

Chief executive Zohaib Ahmed said "Disclosure should be a shared responsibility between the advertiser and the platform hosting the content," and that until a clear federal rule exists brands should "watermark and clearly label AI-generated content."

Why it matters for digital humans: The 46% figure is the one that bears on this industry, because it places synthetic people, not just synthetic backdrops, inside mainstream brand advertising at a scale nobody has disclosed. It should be read with care: Resemble AI sells synthetic-media detection, the report is a marketing asset, and the detector's false-positive and false-negative rates are not published, so "flagged as AI" is not the same as confirmed synthetic. The disclosure number is the more durable finding. Whatever the true prevalence, labelling is effectively absent and C2PA provenance is absent entirely, and that is precisely the gap that synthetic-performer disclosure rules and EU AI Act Article 50 transparency duties are written to close. Teams shipping avatar-based ad creative should assume those obligations will be enforced against them before platforms build the tooling to make compliance easy.