Just over a year after the public debut of ChatGPT, aforethought about its impact on the internet is becoming a reality: the bulk of AI spam is escalating. In the past week, three instances highlighted the tangible effects of this trend.
First, a recent tech blog called 404 Media had to change its website because of an overwhelming amount of spam generated by artificial intelligence.
The blog observed AI-generated versions of its exclusive content appearing on spam websites, strategically designed for search engine optimization. These spam pages occasionally outranked authentic 404 Media articles in Google search results. The perpetrators profit by displaying ads on these AI-generated pages.
In 404 Media’s exploration of the article-theft geography:
In recent weeks, Jason has delved into researching and experimenting with various AI tools that claim to “spin” articles for users. One such tool, SpinRewriter, empowers users to generate 1,000 slightly different versions of the same article with a single click. It allows automatic publication to numerous WordPress sites through a paid plugin and includes a feature for managing multiple websites from a unified dashboard.
Another company, Byword, boldly promotes the “SEO heist” that successfully “appropriated 3.6M total traffic from a competitor” using a seemingly unconventional method. It involved exporting the competitor’s sitemap and employing AI to create versions of 1,800 of their articles.
The impact of AI-generated article versions extends beyond the theft of clicks and revenue, directly harming the news industry. This practice undermines the efforts of honest platforms that invest significant time and resources in journalistic reporting.
Additionally, Wired reported an unsettling incident involving The Hairpin, a once-popular indie blog from the 2010s. It fell victim to an AI click farmer who, while preserving some popular articles, replaced female authors’ names with male names—an objectionable practice.
At the most obvious end of the AI-spam spectrum, there are instances of AI-generated obits loaded with errors, causing genuine distress to grieving families. As early as 2021, Wired highlighted the activities of “obit pirates” who scraped funeral home websites. Now, leveraging AI, they employ a new and lucrative tactic—creating YouTube videos and spammy websites from obits. It not only capitalizes on search traffic for those seeking information about the deceased but also adds a distressing layer to an already sensitive situation.
The New York Times recently highlighted the suffering caused to a grieving family due to AI-generated YouTube videos. Following the accidental death of a college student on New York subway tracks, scammers swiftly exploited the heightened search interest around the young man’s name and the term “subway.”
In response, these scammers utilized vital terms, instructing AI to produce an obituary in a conversational tone and posting it on a website. Despite containing inaccuracies, the site appeared in Google searches, adding to the distress of the mourning family.
The Hairpin’s intrusion, the obituary pirates, and the 404 Media imitations all involve malicious actors, scammers, and spammers who use artificial intelligence (AI) to churn out vast amounts of content to manipulate Google search algorithms for financial gain.
Beyond the challenges journalists face tussling with content theft and grieving families rightfully distressed by digital exploitation, the issue extends to a significant concern for Google. This predicament delivers subpar results to users, who now have increasingly appealing alternatives—also powered by AI—for their search needs.
Google informed The New York Times that it is aware of spam obituaries and is actively working to address them, having already taken down some that violated its policies.
However, staying one step ahead of these challenges proves challenging for platforms, exemplified by the recent increase of AI-generated inappropriate images of Taylor Swift on X. As AI continues to reshape the internet, both for better and worse, the responsibility is on Google and the companies behind these AI tools to mitigate the actual harm imposed by such occurrences.
