Resemble AI, a Mountain View-based synthetic-media detection firm, released a study on Oct. 22, 2026 that scanned 37,768 Facebook and Instagram ads and flagged 3,379 as AI-generated or altered. Only two of those flagged files mentioned AI in the copy, a shortfall that matters as New York and the EU now require clear disclosure of synthetic content.
What happened
Resemble AI examined 37,768 image and video files from Meta’s public ad archive collected in September 2026. Its detection tool, Resemble Detect, labeled 3,379 files, 8.9% of the total, as AI-generated or manipulated. The audit covered 509 companies across 26 industries, including Temu, Wayfair, Lufthansa, Teladoc and Turkish Airlines. Of the 3,379 flagged files, just two mentioned AI anywhere in the ad copy.
The report, titled the 2026 Synthetic Advertising Report, found that 70% of the fakes were photoreal, built to pass as genuine photos or videos, and 46% featured a person. None of the 971 files checked for provenance metadata carried any disclosure. The study also identified a second collection of betting and crypto ads where audio analysis caught additional fakes, noting that "Every rate is a floor."
Why it matters
The timing aligns with two new legal requirements. New York’s synthetic performer disclosure law took effect on June 9, 2026, and Article 50 of the EU AI Act became enforceable on August 2, 2026. Both statutes place the disclosure duty on the creator or deployer of the ad, not the platform. Because the vast majority of AI-generated media in Meta ads lack any textual acknowledgment, brands risk non-compliance, potential fines, and reputational damage.
The audit also revealed that disclosure rates differ sharply by page type. On brands’ own pages, 1,971 of 28,463 files (6.9%) were flagged, and 57% of the 389 brands examined ran at least one AI file. By contrast, “bait” pages, sites that use brand names to lure clicks, showed a 71.3% flag rate. This disparity suggests that third-party advertisers are more likely to deploy synthetic media without disclosure, increasing the compliance burden for brand owners who must monitor not only their own pages but also external affiliates.
What this means for marketers
Marketers should treat AI-generated content as a compliance liability, not just a creative tool. First, audit all ad assets, including those run by partners, affiliates and user-generated content channels. Use detection services similar to Resemble Detect to flag synthetic media before publishing. Second, embed clear disclosure language wherever AI is used, satisfying both New York and EU requirements. Third, consider adopting a verified lead generation approach, such as ADFLU’s pre-qualified leads service, to ensure that any AI-enhanced creative aligns with trustworthy data practices and reduces the risk of fraud.
Finally, build internal policies that assign responsibility for AI disclosure to the campaign owner, mirroring the legal focus on the creator or deployer. Regular training on emerging AI regulations will keep teams ahead of enforcement actions.
The bottom line
Resemble AI’s large-scale audit shows that synthetic media is already woven into a noticeable share of Meta ads, yet disclosure is virtually nonexistent. With new laws in New York and the EU demanding transparency, brands that ignore the hidden AI in their ad ecosystems face legal and reputational exposure. Proactive detection, clear labeling, and strict governance are essential steps for marketers to stay compliant and maintain consumer trust.
Sources
- 1.ppc.land