More than 2,100 AI-slop websites found monetizing content through programmatic ads

A new investigation by advertising supply chain intelligence company DecryptAds has identified 2,116 AI-generated recipe websites operating within the programmatic advertising ecosystem, raising concerns about advertising transparency, publisher verification and the growing volume of low-quality inventory available to advertisers.

The research found networks of websites using fabricated author identities, AI-generated content and extensive advertising placements to generate revenue. According to DecryptAds, ads from major brands, including Adobe, Capital One, Pepsi, LinkedIn, Siemens and Google Workspace, appeared across the identified websites.

The findings highlight an emerging challenge for the digital advertising industry as generative AI makes it easier to create large networks of content-driven websites that rely on automated advertising systems for monetization.

The websites identified in the investigation largely focus on recipes and food-related content, although researchers found connections to other categories, including political news, automotive content and gaming forums.

Many of the sites follow practices associated with made-for-advertising (MFA) publishers, which prioritize advertising impressions and revenue generation over original editorial content or user experience.

According to DecryptAds, the investigated websites frequently feature multiple advertising placements alongside recipes generated or assembled using AI tools.

Some reproduce popular food trends or publish unusual recipe combinations that appear designed to attract search traffic rather than provide tested cooking instructions.

The research also identified websites using fabricated author biographies and AI-generated profile photographs to present their content as the work of real individuals. In one example, a website used imagery resembling actress Emily Blunt to represent a fictional recipe creator named Elena. Other websites reportedly used stock photography, manipulated images and invented personal histories to establish credibility with visitors.

These practices raise questions about content authenticity and whether advertisers have sufficient visibility into the websites receiving their programmatic advertising budgets.

DecryptAds began its investigation by examining advertising technology companies operating in jurisdictions it classifies as presenting elevated geographic or regulatory risks.

The researchers analyzed relationships involving approximately 300 advertising technology companies, using publicly available ads.txt and sellers.json records to identify connections between publishers, advertising intermediaries and supply-side platforms (SSPs).

After narrowing the findings to approximately 400 suspicious recipe websites, the team conducted manual reviews of author identities and site characteristics.

Researchers subsequently developed more than 190 technical fingerprints to identify additional websites associated with the same publishing networks.

The process resulted in a dataset of 2,116 websites connected through advertising infrastructure, technical characteristics or apparent ownership relationships.

DecryptAds reported that some websites had advertising connections involving companies based in Russia, China, Turkey, Cyprus and the United Arab Emirates.

The company categorizes certain jurisdictions as higher-risk based on factors including sanctions exposure, cybersecurity concerns and regulatory conditions.

However, the investigation did not establish that the websites or their advertising partners had committed cybercrimes, and the geographic classification does not independently demonstrate unlawful activity.

The findings instead point to the complexity of programmatic advertising supply chains, where multiple intermediaries can make it difficult to establish which companies are directly responsible for monetizing individual websites.

The investigation identified several advertising platforms appearing in the websites’ publicly available monetization records. DecryptAds reported connections involving HBAgency on 1,631 websites, Ezoic on 407 and MonetizeMore on 11.

However, companies contacted by AdExchanger disputed aspects of the findings or emphasized that publicly listed advertising relationships do not necessarily demonstrate active monetization.

HBAgency said it maintains publisher screening procedures, monitors invalid traffic and does not work with partners in sanctioned jurisdictions. It also challenged the report’s interpretation of certain indirect advertising relationships.

MonetizeMore said it was not actively monetizing any of the 11 websites identified in the research and had initiated a broader review of its inventory.

Ezoic argued that AI-generated content should be assessed according to its accuracy, usefulness and audience engagement rather than its production method alone. The company said it rejects more than 98% of websites applying to its platform.

Mintegral also disputed an implied connection to one of the investigated websites, stating that its records showed no direct advertising activity on the domain.

These responses underline the limitations of ads.txt and sellers.json records, which identify authorized advertising relationships but do not always establish whether advertising transactions are actively taking place.

The investigation coincides with updated Google Search guidance addressing websites that use fictional creator identities to misrepresent the origins of their content. Google’s guidance identifies fabricated author profiles, including AI-generated headshots and invented professional credentials, as deceptive practices that can indicate low-quality content.

However, the guidance applies to Google Search quality assessments and does not automatically establish advertising eligibility rules across independent programmatic exchanges.

The distinction is relevant as advertising platforms consider how to evaluate AI-generated publisher inventory.

The research raises questions about whether existing brand safety tools, publisher verification procedures and supply-chain transparency standards can adequately identify large networks of automatically generated websites.

For advertisers, the findings highlight the potential for programmatic budgets to reach websites with misleading authorship claims and limited editorial oversight.

DecryptAds has called for closer scrutiny of publisher identities and advertising intermediaries, while acknowledging that attribution across complex reseller relationships remains difficult.

The report does not quantify the total advertising revenue generated by the identified websites or establish the amount of advertiser spending affected.