Resemble AI finds AI media in 1 in 11 Meta ads

Resemble AI’s audit of 37,768 ad files from 509 brands found that AI-generated or manipulated media already appears in 8.9% of ads on Meta platforms, with most synthetic ads designed to look real and almost none disclosed. The report raises fresh questions about consumer trust, consent and compliance as AI labeling rules expand globally.

Why it matters: - AI-generated advertising is no longer niche on Facebook and Instagram. - The study found a disclosure gap that can mislead consumers, complicate consent and increase legal risk for brands. - Real creators, athletes and customers can be impersonated without knowing their likeness or content has been used.

What happened: - Resemble AI released its 2026 Synthetic Advertising Report on Oct. 22, 2026. - The company scanned 37,768 image and video files from 15,433 ads run by 509 companies across 26 industries. - The audit covered ads on Meta platforms, including Facebook and Instagram. - The study found 8.9% of ad files contained AI-generated or manipulated media. - More than half of the brands scanned, 222 of 389, had run at least one AI ad from their own page. - Among brands that used AI more than once, one in eight of their ads was synthetic.

The details: - The most common synthetic ads were built to look real. - 70% of AI ads were photoreal, meaning they were designed to pass as a real photo or real video. - Nearly half, 46%, depicted a person. - Only about 1 in 9 fakes was stylized enough that a viewer would likely recognize it as AI. - Adoption appeared across every industry, including consumer packaged goods at 13.4%, consumer electronics at 12.6%, banking, pharma and fast food. - Temu, Wayfair, Motel 6 and Lufthansa used AI in more than half of their sampled ads. - Brands used AI to create selfie-style testimonials and phone-shot videos that mimic real customer footage. - Resemble Detect flagged UGC-style ads from Teladoc, MoneyLion, Dave and Tarte Cosmetics as AI-generated with 99% or higher confidence. - Just 2 of 3,379 AI files, or 0.06%, mentioned AI in the ad copy. - Only one brand, Turkish Airlines, proactively disclosed its own ad. - None of 971 files checked carried C2PA content credentials, including every confirmed fake. - In a separate scan of nine sportsbook, prediction market and crypto brands, 44% of fakes, or 62 of 142, were caught only by audio. - Those ads paired real footage with a synthetic voice. - Crypto and brokerage ads were the most synthetic category in the study at 29.3%. - OKX ran AI in nearly half its ads. - YouTuber Elliot Choy found a Kalshi ad built from his apartment-move video, with AI swapping him for a white man who does not exist.

Between the lines: - The report argues that the main risk is not simply that brands use AI, but that they use it without labeling it. - The same tools that help marketers polish creative can also help scammers imitate brands and people. - New York’s synthetic performer disclosure law took effect in June. - South Korea requires on-screen labels for AI-generated virtual people in ads. - The EU AI Act’s labeling rules are also coming into force. - Most of these rules place the burden on advertisers rather than platforms. - That shifts compliance, monitoring and disclosure responsibility closer to brands and agencies.

What’s next: - Resemble AI is urging marketers and agencies to audit where synthetic media enters their creative pipeline. - The company recommends treating synthetic people differently from synthetic backgrounds or music. - Resemble AI also recommends disclosing AI use by default, watermarking brand-made content and monitoring for fake ads using a brand’s name. - The company says the full report, methodology and calculation appendix are available here.

The bottom line: - AI ads are already mainstream on Meta, but labeling is still almost nonexistent. That gap is becoming a trust issue for consumers and a compliance issue for brands.