AI-Generated Spam Outranks Genuine Journalism in Google News, Exposing Flaws in Search Algorithms

To review this article again, navigate to My Profile and then select Saved stories.

Google's Search Algorithm Prioritizes AI-Generated Content Over Genuine Journalism in News Searches

While browsing Google recently, I encountered an article that strangely seemed like something I had seen before.

In my quest to find the most current details about Adobe's stance on artificial intelligence, I entered "adobe train ai content" into Google's search bar and navigated to the News section. I was already familiar with the WIRED article that was listed second on the results page, titled “Adobe Says It Won’t Train AI Using Artists’ Work. Creatives Aren’t Convinced.” Additionally, while the name of the top-listed publication, Syrus #Blog, didn't ring any bells for me, the title of their article immediately felt familiar: “When Adobe promised not to train AI on artists’ content, the creative community reacted with skepticism.”

Upon selecting the first link, I was redirected to a cluttered site filled with content that was clearly copied, featuring numerous articles that had been slightly altered, often accompanied by illustrations created by artificial intelligence. Within this dubious article, the entirety of the WIRED publication had been replicated, with minimal adjustments made to the wording. The original quotations were also taken without permission. The sole attempt at giving credit was a solitary link at the end of the page, pointing to our original article.

Google's search results display a selection of news pieces, placing an AI-generated spam variant of a WIRED article prominently at the top, while the authentic report is positioned right below it in second place.

The artificial intelligence wasn't merely replicating news articles in English; my investigation revealed identical instances of this pilfered material in 10 different languages, covering several of the languages that WIRED publishes in, such as Japanese and Spanish.

Content initially released in platforms such as Reuters and TechCrunch was copied onto this blog in various languages, accompanied by comparable AI-generated images. In the period of late June to early July, as I was investigating this matter, it seemed that the site Syrus had successfully manipulated Google's News search results to appear on the first page for several technology-oriented searches.

For instance, when I looked up “competing visions google openai,” I found a TechCrunch article leading the Google News results. Following it, pieces from The Atlantic and Bloomberg were listed, analyzing the contrasting strategies of the competing AI firms. However, positioned just beneath these more credible sources was another entry from Syrus #Blog, which was notably derivative of the TechCrunch article at the very top.

In January, 404 Media reported that at the start of the year, AI-generated content was frequently showing up in response to simple searches on Google News. Two months following this, Google unveiled a series of significant updates to its search algorithm and introduced new measures to combat spam, aiming to enhance the quality of its search results. By the end of April, Google announced it had completed major updates aimed at eliminating ineffective content from its search rankings. "We have fully implemented these updates as of April 19. As a result, there's a 45 percent reduction in search results featuring poor-quality, non-original content, surpassing our initial expectation of a 40 percent improvement," Elizabeth Tucker, a Google product management director, stated in a company blog post.

Despite modifications, the problem of spam content generated through artificial intelligence continues to be a persistent challenge for Google News.

"Currently, Google is grappling with a widespread issue, the exact reasons for which are difficult to pinpoint," explains Lily Ray, who occupies the role of senior director of SEO at Amsive, a marketing firm. "Several of our clients have brought to our attention instances where their original articles were essentially copied and then altered through AI, resulting in content that closely mimics their own work but appears as a scrambled, AI-modified reproduction of it."

Upon initial observation, it became apparent that a number of the visuals accompanying Syrus' blog posts were produced by artificial intelligence. This conclusion was drawn from the noticeable anomalies such as sagging eyes and various distorted body parts, which are common indicators of AI's attempts at depicting humans.

Did the content of our piece undergo a transformation through artificial intelligence? To uncover the truth, I made contact with the individual responsible for the blog to gain deeper insight into its creation. Through email correspondence, I got a reply confirming that an Italian marketing firm was the architect behind the blog. This firm acknowledges utilizing an AI application during the drafting phase. "In response to your inquiries about copyright issues, we want to reassure you that our methodology for producing content leverages AI technology to gather and reprocess information from a multitude of sources, all the while adhering to the principles of copyright law," is the explanation provided by someone identifying themselves as Daniele Syrus in an email.

Critics argue that the mere presence of a solitary hyperlink at the article's end does not constitute proper credit. Merely including a link, without naming the source, falls short of being a valid argument against allegations of content theft. Additionally, the individual maintains that the aim of the website isn't to attract traffic via Google searches but rather to experiment with artificial intelligence algorithms across various languages.

In response to an email inquiry, Google chose not to provide specific remarks concerning Syrus. Meghann Farnsworth, representing Google, stated, "Our stance is not to discuss individual websites. However, under our revised spam guidelines, we strictly forbid the production of mass, derivative content aimed at achieving high search engine rankings." She added, "We globally enforce measures against websites that violate our guidelines." (It's noted that Farnsworth previously worked for WIRED.)

Upon examining Google's policies on spam, it's evident that this blog is in breach of the company's regulations against web scraping. The policies cite abusive scraping practices as including websites that replicate content from other sources, make minimal alterations (such as changing words for synonyms or employing automated methods), and then redistribute it. Farnsworth did not verify if this blog contravenes Google's guidelines or if Google intends to lower its ranking in Google News search results following these findings.

How can authors of original content ensure their work is adequately safeguarded? The answer isn't straightforward. However, after numerous discussions with SEO professionals, a prominent theme has emerged: a pervasive feeling of unease.

"Andrew Boyd, a consultant for the digital link-building agency Forte Analytica, highlights a significant problem in our industry, noting, "We're essentially dealing with a form of trauma, and that's not an exaggeration." He believes a key issue is the lack of options available to publishers who find themselves adversely impacted. "Imagine starting your day only to discover you've lost half of your audience overnight," Boyd illustrates. He points out that several websites have seen their visitor numbers plummet following updates to Google’s search algorithm over time.

Numerous SEO professionals express frustration over Google's opacity regarding significant updates. However, not all the individuals I conversed with were disapproving of the spam that appears in search outcomes. "In reality, Google's most formidable obstacle is combating spam, a fact for which it seldom receives praise," observes Eli Schwartz, who penned "Product-Led SEO." "Thus, despite the ongoing grievances concerning Google's current performance, a search for something like hardware won't mistakenly lead you to adult content. They’re managing quite well in that aspect." To counter spam, the company is consistently rolling out minor updates to its search engine.

Certainly, Google occasionally provides a satisfactory service by shielding users from encountering dubious adult content during searches for mainstream topics. However, it's fair to demand more from one of the globe's most dominant corporations—a company with significant sway over the creation, dissemination, and consumption of digital content—in terms of more effectively eliminating copied and unconstructive content from its News section.

"Ray expresses his frustration, noting the challenges they face as they strive to maintain high standards, only to be overshadowed by the surge of inferior AI-generated content. "It's disheartening to witness the impact this is having, breeding unprecedented levels of tension and hostility within our industry, something I haven't encountered in my 15 years of experience," he says. Ray warns that if search engines continue to favor these low-effort AI sites, it could discourage publishers from investing in quality content, ultimately eroding users' confidence in the reliability of top-ranked sites on Google News."

Explore More Options…

Dive into election season with our WIRED Politics Lab newsletter and podcast.

Unconvinced that breakdancing qualifies as an Olympic event? The global champion shares your sentiment (to some extent).

Investigators deciphered the password to a cryptocurrency wallet valued at $3 million after 11 years of it being locked.

The remarkable emergence of the world's inaugural AI beauty contest

Ease the strain on your spine: Discover the top office chairs we've evaluated.

WIRED PROMOTIONS

Additional 20% Discount on Selected Dyson Products for Owner Rewards Members

Enjoy Up To An Additional 45% Discount During The June Sale

Promotional Code for Vista Print: Save 20% on Specific Signs

Discount Code for 10% Reduction at Newegg

Subscribe to Peacock Premium for just $49.99 annually, which breaks down to $4 per month.

Discover DJI's Exclusive Discounts and Deals for Students in 2024

Additional Content from WIRED

Evaluations and Tutorials

© 2024 Condé Nast. Rights reserved. WIRED might receive a share of revenue from items bought via our website, thanks to our Affiliate Relationships with retail partners. Content on this website is protected and cannot be copied, shared, broadcast, stored, or used in any manner without explicit approval from Condé Nast. Advertising Choices.

Choose a global website