March 2024: Sites Penalized in Google’s AI Content Update

March 2024: Sites Penalized in Google's AI Content Update

Google declared on March 5th that a significant update would be rolling out to remove 40% of irrelevant content. The first step in this update was to fully deindex all sites through a substantial round of manual actions. The algorithm update, which will be released in two to four weeks, is different from the manual actions that are taking place in tandem with this update.

To help people understand these updates, we have finished two major studies in the last 24 hours, both of which are included here. 

Study 1: List of Websites That Had a Manual Action in March 2024 

Out of 79k websites checked in March 2024, 4,446 websites had a manual action applied to them.

Study 2: Was AI Content Spam to Blame for the Manual Action?

Fifty percent of the websites had between ninety and one hundred percent of their posts generated by artificial intelligence (AI).

This post will go over the following topics:

  • An overview of the Google update
  • What manual action and site deindexation are
  • The fullest list of affected websites; and
  • What we’ve discovered about those affected websites.

An overview of the AI Content Manual Action Update from March 2024

Google released this update on March 5th, to penalize

 “Scaled content abuse” 

Putting an end to the mass publication of content spam (after all, who uses AI in 2024 and beyond to create content spam?).

 “Expired Domain Abuse” 

Aims to prevent individuals from using expired domains for search engine optimization.

“Site Reputation Abuse” 

Aims to prevent trustworthy websites from tricking search engines by publishing articles about parasitic search engine optimization on a portion of their website. The Sports Illustrated Parasite SEO example serves as an illustration of this.

This page provides coverage of Google’s March 2024 core update and multiple spam updates:

Google Blog Posts:

It is important to understand that human activities are not the same as the algorithm changes that will take place. 

What is a “Manual Action” and “Site Deindexation”? 

If Google finds a website that violates its policies, it may take manual action to remove it entirely from its search results, also known as “deindexing” the website. On March 5th, sites began to receive more and more of these notifications in their Google Search Console Manual Action dashboard.

It looks like these manual actions have the effect of completely removing the content from Google’s search results.

What information do we currently have about the affected websites?

Study1: List of Websites That Had a Manual Action in March 2024

We have finished a study to identify content websites that Google deindexed but had Google Organic traffic until very recently to better understand the scope of this manual action effort. This study is not focused on e-commerce or other types of websites, but rather on content-first websites.

Findings:

  • Among the approximately 79k sites analyzed, 1.9 per cent had a manual action applied to them. 
  • The cumulative traffic loss was estimated to be over 20 million visitors per month.
  • Three websites with over one million organic visitors per month were reduced to zero. Manual action was applied to over 1,446 sites that were on MediaVine, Raptive, or Ezoic.

3 Websites With More Than 1 Million Organic Visitors Per Month to Zero

  • zacjohnson.com
  • beingselfish.in
  • equityatlas.org

Methodology

This is how the study was carried out.

Summary: We compiled a list of 79k websites that represent the internet better than average. We then verified whether or not these websites are currently indexed, and if not, we looked up their recent organic traffic history using two different sources: AHrefs and SimilarWeb. 

In case a website experienced organic traffic in February but is currently not indexed by Google, a manual action was likely applied on March 5th.

1. Using BuiltWith lists, we generated a list of all URLs displaying ads from well-known advertising providers. AdSense sites were not examined because of the large number of low-quality sites; instead, these platforms were chosen because they have minimum requirements before allowing sites to be added. This list may underestimate the percentage of websites that were subject to manual action, but it should (arguably) be a better-than-average representation of all websites on the internet. According to the company supplying the advertising, the following websites were counted: 

  • MediaVine: 21,808; 
  • Raptive: 6,428; 
  • Ezoic: 51,293.

2. 18 duplicates were eliminated. 

3. We used AHrefs and SimilarWeb to check February Organic traffic numbers for every deindexed site. 

4. We also checked each URL to see if it had been deindexed by searching for “Site: URL.”.

5. We verified that our approach successfully captured a portion of the publicly accessible websites, which it did.

Study 2: Was the Manual Action Caused by AI Content Spam?

Numerous media outlets hastily concluded that the purpose of this update was to eliminate AI spam from Google’s search results. 

While many SEOs on X agree..

 We wanted to conduct a more thorough analysis using our AI Checker.

To determine the prevalence of AI Content on the sites that were manually penalized, we examined 100 recent articles that have already been shared publicly for each of the deindexed sites. ‍

Findings: 

  • 100% of websites with manual action showed AI use.
  • Seven of the fourteen websites we examined had more than 90% of their sample articles. 

Methodology

1. Identified websites that were deindexed and whose URLs had already been made public.

  • Xfresherslive.com
  • qmunicatemagazine.com
  • hnbgu.net
  • zacjohnson.com
  • newsunzip.com
  • Bognor.newspopularbio.com
  • popularbio.com
  • popularnetworth.com
  • bioofy.com
  • istaunch.com
  • healthyceleb.com
  • GoDownSize.com
  • networthpost.org
  • tvguidetime.com
  • thesocialtalks.com
  • juliangoldie.com
  • chipperbird.com
  • EquityAtlas.org
  • filmifeed.com

2. Scraped 100 of the most recent announcements that were more than 100 words long.

3. Several websites were excluded:

  • juliangoldie.com – no longer fully deindexed – 10 pages are in the index (but publicly admits to using AI)
  • chipperbird.com – unable to get content (but the website owner publicly admits to using AI)
  • equityatlas.org – unable to get content
  • filmifeed.com – unable to get content
  • thesocialtalks.com – unable to get content

4. On March 7, we used the model 2.0 Standard to run each article through our AI detector (detector efficacy).

5.  We then finished an analysis for each site to determine the average AI score and the percentage of articles that were suspected of being artificial intelligence (AI)-generated.

Google faces an existential threat from the possibility of an overabundance of spam generated by artificial intelligence. This appears to be a direct attempt by Google to express its opinions about AI-generated spam as well as to penalize it.

News Source: originality.ai

Leave a Reply

Your email address will not be published. Required fields are marked *