The advertising industry’s long-held apprehension that generative artificial intelligence would catalyze an unprecedented surge in "made-for-advertising" (MFA) websites, designed solely to exploit programmatic advertising revenue, is rapidly solidifying into a concerning reality. Recent research has begun to illuminate the extent of this phenomenon, with a particular focus on AI-generated recipe sites that mirror the tactics of established MFA operations. These platforms, characterized by an overabundance of ad placements and low-effort content, are now attracting the attention of major brands, raising significant questions about supply chain integrity and cybersecurity.
The Rise of AI-Generated Content Farms
The proliferation of AI-generated content, particularly in the form of recipe websites, has been a growing concern within the advertising technology ecosystem. Industry insiders have consistently warned about the potential for AI to lower the barrier to entry for creating vast networks of low-quality websites designed to monetize through programmatic ad tech. This has been termed "AI slop," a digital byproduct of generative AI that floods the internet with often nonsensical or poorly constructed content, optimized not for user engagement but for ad impressions.
A groundbreaking study released by advertising supply chain intelligence company DecryptAds has provided concrete evidence supporting these fears. The research uncovered over 2,100 AI-generated recipe sites that meticulously follow the established MFA playbook. These sites are laden with multiple, overlapping ad placements on pages filled with content that demonstrates a clear lack of human oversight. Disturbingly, the study found that advertisements from prominent blue-chip and Fortune 500 companies are appearing on these AI-generated recipe platforms, indicating a potential blind spot in brand safety protocols.
Unpacking the MFA Playbook in the Age of AI
The DecryptAds report examined not only recipe sites but also highlighted that the publishers behind these operations often run a diverse array of other ad-supported platforms. These include political news aggregators, and forums dedicated to cars, anime, and video games. This diversification suggests a broader strategy of content farming, leveraging AI to populate various niches with ad-heavy, low-value content.
By meticulously analyzing ads.txt and sellers.json files—essential documents for verifying the legitimacy of ad inventory sellers—DecryptAds identified that these recipe sites are primarily monetized by Supply-Side Platforms (SSPs) that collaborate with ad tech partners situated in what DecryptAds categorizes as "risky geographies." Zach Edwards, co-founder and chief of product and research at DecryptAds, explained that these "risky geographies" are regions often associated with a high prevalence of cybercrime or that maintain adversarial relationships with the United States.
Cybersecurity and Data Privacy Concerns

The implications of US citizens, including potentially government officials, accessing these sites are significant. Edwards voiced serious cybersecurity concerns, noting the potential for ad tech vendors to harvest user data and for these platforms to be used for malvertising schemes. While DecryptAds has not yet uncovered definitive evidence of criminal activity directly linked to these sites or their ad tech partners, the inherent risks are substantial. The aggregation of data from unsuspecting users, combined with the potential for malicious ad campaigns, creates a fertile ground for exploitation.
The ad inventory offered by these sites is of dubious value, Edwards asserted. Similar to traditional MFA sites, these AI-generated platforms are densely packed with ad placements engineered to artificially inflate viewability metrics, a common tactic to deceive advertisers into believing their ads are being seen. The content itself often involves viral recipe trends or slightly modified versions of popular recipes from established cooking sites. These AI-generated recipes can be nonsensical, such as the example cited: "Hot Cocoa Buffalo Chicken Gingerbread Dip with Marshmallow Topping," a concoction that suggests an algorithm prioritizing keyword stuffing over culinary sense.
Edwards concluded that such sites have no legitimate place in a healthy programmatic ad supply chain, especially in light of recent changes to Google Search. These updates aim to deprioritize websites that feature AI-generated authors and creators, signaling a broader industry push towards transparency and authenticity. However, the decision of whether to serve ads on such platforms remains a contentious issue for many SSPs, highlighting the ongoing struggle to maintain quality standards in a rapidly evolving digital landscape.
Fabricated Authorship: The Illusion of Authenticity
A significant challenge in policing these AI-generated sites lies in their deliberate design to mimic the appearance of human-created content. The networks of sites identified by DecryptAds feature fabricated author profiles on their "about" pages. These profiles often include AI-generated human avatars, stock images, stolen photographs of real chefs, and even subtly altered images of celebrities, such as actress Emily Blunt. This practice is a direct violation of Google’s recent "helpful content" update, which explicitly states that fabricating creator profiles, whether through AI-generated headshots, made-up names, or false credentials, constitutes deception. Such deception renders a page untrustworthy and indicative of low quality, both for users and automated quality systems.
Edwards emphasized that this Google policy serves as a crucial signal to the advertising industry. Websites that fail to disclose their use of AI-generated authors risk being officially classified as "low-quality," making them unsuitable for monetization. This classification has significant implications for publishers, ad tech vendors, and ultimately, advertisers seeking to reach genuine audiences.
However, a source within an SSP, speaking anonymously to AdExchanger, offered a counterpoint. This individual suggested that Google’s guidance might be specific to its search business and should be considered separately from best practices for Google Ads. This perspective also challenged the notion that AI-generated author profiles inherently devalue content, arguing that real users do engage with AI-generated content, implying that value can still exist if the content itself is useful.
DecryptAds’ Methodological Approach: A Deep Dive into the Supply Chain
The core concern that brought these AI-generated recipe sites to the DecryptAds team’s attention was not solely content quality but the potential for cybercrimes and state-backed malicious actors to exploit the ad supply chain. While the report does not allege direct criminal activity by these sites or their partners, the operators are described as opportunists employing a low-effort content farm strategy rather than sophisticated threat actors.

The cybersecurity concerns directly informed DecryptAds’ methodology. The team’s regular monitoring of ad tech companies registered in "risky geographies"—including Russia, China, Cyprus, Turkey, and the United Arab Emirates—led to the discovery of these recipe sites. These nations were selected due to their associations with high degrees of cybercrime, adversarial relationships with the US, or a lack of robust privacy firewalls between governments and tech companies. DecryptAds categorizes these nations into threat tiers, with US-sanctioned countries like Russia at the highest risk level.
DecryptAds initiated its investigation by examining the supply chains of approximately 300 ad tech companies operating in these risky regions. By scrutinizing ads.txt files, the team identified publishers working with these companies. Further analysis of sellers.json files maintained by SSPs, alongside domain name system (DNS) data, SSL certificates, and Google Analytics data, allowed DecryptAds to determine which SSPs were partnering with these ad tech firms.
Using a curated list of food-related keywords, Edwards narrowed down the list of sites monetized by these ad tech companies to an initial pool of "400 suspicious recipe websites." Manual verification confirmed that each of these websites employed AI-generated avatars while obscuring their true authorship. AdExchanger’s own testing of several active URLs provided by DecryptAds corroborated these findings, with many sites failing AI detection or featuring easily identifiable stock author images.
Subsequently, DecryptAds created over 190 "technical footprints" to identify other sites operated by the same owners. This comprehensive approach led to the discovery of 2,116 such sites that had infiltrated the programmatic ad supply chain through a complex web of reseller relationships among global ad tech partners. These technical footprints also revealed ownership links to additional risky geographies, with many of these sites exhibiting social media profiles connected to Morocco and Algeria.
The intricate nature of the programmatic supply chain, with its numerous intermediaries, creates an environment where shady site operators can easily blend in. Edwards noted that the monetization of these sites suggests that some SSPs and ad tech platforms may not be adequately vetting publishers before onboarding. He acknowledged, however, that even with stringent vetting processes, detecting and eliminating MFA inventory or "AI slop" remains a persistent challenge.
The Imperfect Tools of Transparency: Ads.txt and Sellers.json
The DecryptAds research underscores the critical role of ads.txt and sellers.json in identifying vulnerabilities within the programmatic chain. These publicly available files are currently the primary means of verifying the relationships between publishers and ad tech vendors. However, their effectiveness is hampered by infrequent updates by many publishers and the potential for errors.
A case in point involves dailytastekitchen.com, a site flagged in the DecryptAds report. Its ads.txt file listed HBAgency as the SSP and indicated connections to eight ad tech companies on DecryptAds’ "risky geo list," including Between Digital (Russia), Mintegral (China), AdMatic (Turkey), and Monetix Ads (UAE), along with four other partners in Cyprus or Turkey.
HBAgency, however, disputed this broad association. A spokesperson clarified that HBAgency directly works only with Monetix Ads. While Monetix Ads reportedly utilizes technology from AdMatic, HBAgency maintained no direct commercial relationship with AdMatic. The spokesperson further emphasized that Turkey, where both Monetix Ads and AdMatic are based, is a NATO member and EU candidate country not subject to sanctions. They asserted that Monetix Ads underwent HBAgency’s vetting process and operates under contractual data protection obligations consistent with GDPR.

HBAgency also stated that every publisher onboarded to its SSP undergoes screening by an independent third-party fraud provider. The company conducts screenings for invalid traffic and implements "enterprise-grade malvertising protection across our supply." Publishers are continuously reviewed, and accounts violating terms and conditions are disabled. Crucially, HBAgency affirmed that it does not serve ads in sanctioned countries like Russia, Belarus, Iran, and North Korea, nor does it work with publishers or partners based in sanctioned jurisdictions.
Despite HBAgency’s assurances, the DecryptAds report alleges that HBAgency is a monetization partner for a significant portion—1,631 out of 2,116—of the AI slop websites it uncovered. This discrepancy highlights the complexities and potential inaccuracies in data interpretation, leaving it unclear whether the issue lies with DecryptAds’ methodology or HBAgency’s vetting processes. This ambiguity, as Edwards pointed out, exemplifies the significant challenges advertisers and SSPs face in their efforts to avoid monetizing such problematic websites.
Edwards noted that DecryptAds intentionally kept its mentions of SSPs brief in the report, acknowledging the difficulty in attributing issues with complete accuracy and confidence due to the intricate nature of these relationships.
Responses from SSPs and Tech Vendors: A Spectrum of Compliance and Skepticism
DecryptAds aims to raise awareness about the escalating volume of AI-fueled MFA inventory, warning that the industry risks being overwhelmed by low-quality sites if proactive measures are not implemented to redirect budgets.
In response to inquiries from AdExchanger regarding their quality control measures, several SSPs and ad tech vendors mentioned in the DecryptAds report provided statements. HBAgency’s response has been detailed above.
MonetizeMore, identified by DecryptAds as monetizing 11 of the AI slop sites, stated that it was not actively running its demand or demand partners on any of the listed sites. The company initiated a broader review of its publisher inventory and committed to demonetizing any sites that fall below its standards. CEO and founder Kean Graham confirmed that one flagged site, Bakerias.com, maintained a legacy connection to their reporting platform but operated its own AdSense account and was not monetized by MonetizeMore due to failing traffic and ad inventory quality standards. Graham also aligned MonetizeMore’s publisher standards with Google’s helpful content update, emphasizing that deceptive authorship violates Google’s guidance and that advertisers should not be purchasing such inventory.
Ezoic, which DecryptAds claims monetizes 407 of the AI recipe sites, contested the notion that AI authorship automatically equates to low-value content. The company asserted that content should be judged on its quality and usefulness, regardless of its production method. Ezoic highlighted its rigorous review process, rejecting over 98% of applicant sites, and emphasized its evaluation of content quality, audience engagement, and real-time brand safety analysis. They argued that AI-generated recipes can provide genuine value to users and advertisers, provided the recipes are accurate and functional.
Culture G and Between Digital did not respond to requests for comment.

Data Security in "Risky Geographies": Mintegral’s Perspective
Mintegral, another company contacted by AdExchanger, focused its response on the data security concerns raised in the report. While listed as an ad tech partner for dailytastekitchen.com by HBAgency, a Mintegral spokesperson stated that their records showed no direct ad serving on that specific site. They reiterated a commitment against fabricated author identities intended to mislead audiences or advertisers.
The spokesperson clarified that Mintegral’s core business is mobile in-app advertising, with limited mobile web inventory accessed through third-party exchanges. They emphasized that bid data received from exchange partners is subject to applicable data protection and privacy safeguards, designed to comply with US privacy law. Regarding concerns about Mintegral being based in China, the spokesperson asserted that the company has no history of security incidents or regulatory actions concerning user data and has not received government requests for data access. They stressed the importance of distinguishing legitimate data security questions from generalized assumptions based on geographic origin.
The Path Forward: A Call for Enhanced Vigilance
The proliferation of AI-generated content farms, as evidenced by the DecryptAds report, poses a significant threat to the integrity of the digital advertising ecosystem. The sophisticated tactics employed by these operators, coupled with the inherent complexities of the programmatic supply chain, demand increased vigilance from all stakeholders. As AI technology continues to advance, the industry must adapt by strengthening vetting processes, demanding greater transparency from publishers and ad tech partners, and actively working to identify and eliminate "AI slop" before it further erodes advertiser confidence and user trust. The ongoing challenge lies in striking a balance between leveraging AI’s potential for content creation and preventing its misuse to create deceptive and low-value advertising environments.






