A NewsGuard report highlights the growing use of AI chatbots to create text on websites, attracting advertisers who unknowingly place ads on such AI-generated sites. Google, despite its anti-practice policies, serves 90% of these ads. This contributes to an internet filled with low-quality AI content and wasted ad money. Programmatic advertising, driven by algorithms, results in brands supporting content farms with clickbait and autoplay videos. Generative AI automates this process, leading to an increase in junk sites. NewsGuard has identified 217 such sites in multiple languages. Although not all spread misinformation, the economic incentives could exacerbate this issue. Policymakers are called to address misinformation while maintaining the internet economy.
A recent report from NewsGuard, a media research organization, has revealed that numerous websites are using AI chatbots to generate text, attracting advertisers who pay for their ads to be displayed on these sites. Over 140 major brands have been found advertising on these AI-generated sites, often unknowingly. Google, despite its policies against such practices, has been found to serve 90% of these ads.
This practice is contributing to an internet landscape that is increasingly filled with low-quality, AI-generated content, and is also leading to substantial wastage of ad money. Most online advertising is done through a process called "programmatic advertising", where algorithms automatically bid for ad spots on various websites based on complex calculations aimed at maximizing visibility. This often results in brands advertising on unknown websites with little human oversight.
Content farms, which produce low-quality content to attract ad revenue, are taking advantage of this situation. These "made for advertising" sites use tactics like clickbait, autoplay videos, and pop-up ads to maximize revenue. The Association of National Advertisers estimates that around $13 billion is wasted globally on these sites each year.
Generative AI is now being used to automate the content farm process, leading to an increase in the number of these junk sites. NewsGuard has identified 217 such AI-generated sites in 13 languages since it started tracking the phenomenon in April. Some of these sites are sophisticated, featuring AI-generated photos and bios of non-existent authors.
NewsGuard identifies these AI-generated sites by looking for error messages typical of generative AI systems. Once these are flagged, a human analyst reviews them. The problem is growing rapidly, with around 25 new AI-generated sites being discovered each week.
Programmatic advertising is the primary revenue source for these AI-generated websites. Many Fortune 500 companies and well-known brands are unknowingly supporting these sites by advertising on them. Google's programmatic ad product, Google Ads, which made $168 billion in advertising revenue last year, has been criticized for serving ads on content farms, despite its policies against such practices.
While most AI-generated sites are considered "low quality", they do not necessarily spread misinformation. However, the economic incentives of content farms could exacerbate the misinformation problem. For instance, one AI-generated site, MedicalOutline.com, was found to be spreading harmful health misinformation.
There are currently no easy solutions to this problem, especially considering that advertising is a fundamental enabler of the internet economy. Policymakers are urged to focus on developing robust mechanisms to prevent the spread of misinformation, rather than banning programmatic ads altogether.
