AI-Generated Clickbait Networks Exploit Reputable Media Outlets, Raises Concerns of Misinformation
In the vast world of the internet, it’s not uncommon to come across junk content and misleading information. But what happens when reputable media outlets become victims of exploitation by AI-generated clickbait networks? Tony Eastin, a former US Air Force member with a background in psychological warfare campaigns, stumbled upon this disturbing phenomenon while researching a pharmaceutical company online. What he found was a small Midwestern newspaper, the Clayton County Register, churning out nonsensical blog posts about retail investing. Intrigued by this discovery, Eastin teamed up with Sandeep Abraham, a former colleague with experience in Army intelligence and the National Security Agency, to dig deeper into the world of AI clickbait.
Their investigation uncovered a network of websites that rely on AI-generated clickbait to deceive audiences and advertisers. These websites, which have acquired URLs once belonging to reputable organizations, thrive on confusion and misinformation. The Clayton County Register, for example, no longer offers local news but has transformed into a financial news content mill. The articles published on the site are not only generated by AI but also feature images created using diffusion models. Some articles even admit their artificial origins, mentioning a company called Automated Insights that specializes in language-generation technology.
Upon examining the bylines on the Register’s former site, Eastin and Abraham discovered that many of the listed writers were not actual journalists. In fact, some shared names with well-known individuals from other fields and had an unrealistically high output. One writer, Emmanuel Ellerbee, credited with numerous articles on bitcoin and banking stocks, shares a name with a former professional football player. The Muck Rack journalist database showed that Ellerbee had published an astonishing number of articles in a short period of time. This raised suspicions about the authenticity of the writers and their content.
The domain of the Clayton County Register changed hands in August 2023, and it was around this time that it began hosting its current financial news churn. Eastin and Abraham discovered that the site attracted most of its readership through search engine optimization (SEO), targeting keywords related to stock purchasing. It also received traffic from social media platforms, particularly crypto news forums on Reddit. The purpose of this scheme was to generate ad revenue from unsuspecting visitors who landed on the site’s garbled content. The algorithmic posts were accompanied by ads served by Google’s ad platform, some of which were unrelated to the content itself.
The identity of the operator behind the Clayton County Register’s former domain remains unknown. However, Eastin and Abraham’s investigation led them to a Polish IT company associated with several domains, including the Clayton County Register. These domains, hosted on a German server, published similar AI-generated content. One of the sites within this network, Aboutxinjiang.com, underwent multiple transformations over time. Initially operated by a Chinese outfit called “the Propaganda Department of the Party Committee of the Xinjiang Uyghur Autonomous Region,” it later became a source of generic financial news posts before redirecting to a page about Polish real estate.
Eastin and Abraham identified nine different websites linked to the Polish IT company, forming an AI clickbait network. These sites were chosen for their preestablished reputations with Google, which helped them gain prominence in search rankings and attract clicks. While Google claims to have systems in place to address attempts to manipulate search rankings, the rise of generative AI tools has made these networks more prevalent. NewsGuard, an online misinformation tracking company, has seen a significant increase in AI-generated content farms within the past year.
The work of Eastin and Abraham sheds light on the operations of AI-generated clickbait networks and the challenges in identifying their operators. These networks often operate anonymously, using services to hide their identities when registering domains. The researchers hope that their findings will encourage people to think critically about the sources of the news they consume and prompt lawmakers to consider measures that can improve the information ecosystem. They continue to investigate other instances of AI-generated content mills and plan to release more reports in the future.
Emerson Brooking, a resident fellow at the Atlantic Council’s Digital Forensic Research Lab, commends Eastin and Abraham’s work for demystifying the use cases of generative AI. While concerns about AI spreading political misinformation are valid, this network demonstrates how content mills focus on uncontroversial topics to generate traffic-based income. It serves as a reminder of how AI is shaping our society, making everything a little more annoying.
The rise of AI-generated clickbait networks exploiting reputable media outlets raises concerns about the spread of misinformation. As technology continues to advance, it becomes increasingly important for individuals to critically evaluate the sources of information they encounter online. By exposing these networks and shedding light on their operations, researchers like Eastin and Abraham play a crucial role in safeguarding the integrity of the information ecosystem.
