Log entry Nº 779 · the stack watch

Bearing NW 315° · October 20, 2026 · 3 min read

Advertising TechNW 315°Watch log

DecryptAds Finds 2,100+ AI-Generated Recipe Sites Feeding Programmatic Ads

DecryptAds found 2,100+ AI-generated recipe sites monetizing programmatic spend. Buyers get a site count, but no spend figures and no named pipes. Here's what that gap means.

By James Calloway3 min read634 words

AI Slop Sites Are On The Rise. Is Ad Tech Ready For Its New MFA Challenge?
Fix Nº 779AI Slop Sites Are On The Rise. Is Ad Tech Ready For Its New MFA Challenge? — kerdkanno / pixabay

Supply chain intelligence firm DecryptAds has uncovered more than 2,100 AI-generated recipe sites designed to monetize programmatic advertising, according to a study released Tuesday. That number matters because it puts hard data behind a warning ad industry insiders have repeated for years: generative AI would supercharge made-for-advertising (MFA) sites.

Now the proof is arriving. And it arrives in a familiar, almost boring category — recipes.

What did DecryptAds actually find?

The study identified more than 2,100 AI-generated recipe sites that, in DecryptAds' framing, take a page straight from the classic MFA playbook. These aren't hobbyist blogs with a chatbot assistant. They're sites built to attract programmatic spend, with content mass-produced cheaply by AI to fill ad slots.

The mechanics are old. The content supply is new. Generative AI has collapsed the cost of producing passable-looking pages to nearly zero, which means the economics that always favored MFA operators now favor them more.

A caveat follows every promise, though. DecryptAds sells supply chain intelligence — the kind of tooling you'd buy if you were worried about exactly this problem. The research is useful, but the messenger has a commercial stake in the alarm. That doesn't make the 2,100 figure wrong. It makes it worth verifying before you brief your CMO.

Why should martech practitioners care?

If you run programmatic campaigns, curate inclusion lists, or sit anywhere near brand safety tooling, this is your problem in three ways:

  • Wasted spend. Every impression served on an AI slop site is budget that bought nothing resembling human attention.
  • Data pollution. Conversions and viewability metrics from MFA inventory corrupt the signals feeding your optimization models.
  • Brand safety drift. Recipe content looks benign to keyword-based filters, which is precisely why the category is attractive to MFA operators.

MFA was never solved. It was managed — partially, unevenly, mostly by large agencies applying their own blocklists. The industry's consistent warnings, which DecryptAds now cites as vindicated, were never matched by consistent enforcement.

Is ad tech actually ready?

Here's what the study doesn't say, and practitioners should notice the gap. DecryptAds counted the sites. It quantified the recipe category. But the published findings don't tell us:

  • How much advertiser spend actually reached these 2,100+ sites, or over what period.
  • Which exchanges or SSPs were selling the inventory, which is the question any serious buyer will ask first.
  • Whether the sites are concentrated with a few operators or spread across thousands of independent domains.

Without those specifics, the number is a symptom, not a diagnosis. It tells us AI slop exists at scale. It doesn't yet tell us whose pipes it flows through.

That distinction matters for accountability. When ANA's 2023-era MFA reporting named the problem, the pressure landed on agencies and DSPs because the supply path was documented. A site count alone gives buyers less to act on — though DecryptAds' core business suggests the full dataset lives behind its paywall.

What should buyers do now?

The immediate move is unglamorous. Audit your active domains against AI-content patterns, not just classic MFA markers like ad density and page-to-ad ratio. Sites that look editorially thin, publish at implausible volume, or cluster around recipe and lifestyle keywords deserve a manual look.

Longer term, the pressure goes to verification vendors and DSPs. If content provenance — knowing whether a machine wrote the page your ad sits on — becomes a standard pre-bid signal, this problem gets tractable. Today, largely, it isn't.

Ad tech has been through the MFA cycle before and largely muddled through. This time the production cost is zero and the volume ceiling is gone. Whether the industry's defenses scale faster than the slop — that question, the study doesn't answer.

Plot points

  1. DecryptAds released its study on Tuesday, identifying more than 2,100 AI-generated recipe sites.
  2. The sites follow the made-for-advertising (MFA) playbook, monetizing through programmatic ad tech.
  3. The findings confirm long-standing industry warnings that generative AI would fuel MFA growth.
  4. The study does not disclose spend levels, implicated exchanges, or operator concentration.

via decryptads.com (Original)

Filed under

  • programmatic-advertising
  • made-for-advertising
  • ai-generated-content
  • brand-safety
  • ad-verification
Share this article:

More from James Calloway

James Calloway

Show full bio

Correspondent covering people, launches and day-to-day practice in marketing technology at Martech Beacon.

31 articles

Neighbouring fixes

‹ Previous articleNext article ›