Brief IA
›

Google targets content abuse, even without AI

AI Tools·Tom Levy·

Google targets content abuse, even without AI

Google targets content abuse, even without AI
Key Takeaways
1Google established a policy against large-scale content abuse in March 2024, with or without automation.
2E-E-A-T appears 120 times in the current guidelines for Google’s quality raters.
3The founder of Algorythmic reports over three years of recovery for a medical site affected by Medic.
Why it matters : human production is not excluded from the scope of large-scale abusive content.

Google established a policy against large-scale content abuse in March 2024, targeting both human production and automation. Its quality guidelines also place significant emphasis on the E-E-A-T framework. The founder of the consulting firm Algorythmic reports the recovery of a medical site after it was rebuilt around experts.

A Medical Site Restored After More Than Three Years

More than three years: this is the recovery timeline reported by the founder of the consulting firm Algorythmic for a client hit by Medic in 2018. The site had developed medical articles using AI and automation, without visible expert qualifications or clearly identifiable human presence. This specialist diagnosed insufficiently demonstrated expertise for health topics falling under the YMYL category.

Initially, the client refused such a significant transformation, and the collaboration ended, reports the founder of Algorythmic. They tried their own methods before returning after nearly two years without recovery. A new strategy was then developed, and the site was fully relaunched about six months later.

The reconstruction relied on doctors, medical reviewers, and their original contributions. According to the founder of Algorythmic, the site regained its previous levels after the launch of this new platform.

Google Regulates Several Forms of Mass Production

The scope of the policy introduced in March 2024 does not depend solely on the tool used: the abuses targeted can stem from individuals, automated processes, or their combination. Generative AI is at the forefront of examples, but machine translation, synonym substitution, and content assembly are also cited. These practices are now part of marketing professionals' experiments with large language models.

These categories have specific precedents. In August 2010, John Mueller from Google explained that a translation published directly from Google Translate could fall under automatically generated content that violates guidelines; he recommended prior human correction. A guide published by Google in 2012 then cited translations without human editorial intervention, Markov chains, automatic synonym replacement, scraped content, and page assemblies without sufficient contribution.

As early as 2003, guidelines for webmasters prioritized user needs over those of search engines and warned against standardized pages that were almost devoid of originality. Google stated that following their spirit would be preferable to seeking loopholes. Several categories of abusive practices have thus traversed over 20 years of rules, with varying formulations.

La lettre du soirGratuit · 19 h

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

In February 2023, just months after the arrival of ChatGPT, Google already stated that it had extensive experience with automation aimed at manipulating its results.

Guidelines Mention E-E-A-T 120 Times

The framework now referred to as E-E-A-T appears 120 times in the current version of the guidelines provided to Google’s quality raters. The company began using its previous acronym, E-A-T, in 2014. However, it is not a unique ranking factor that acts uniformly across all sites; Bing emphasizes quality and credibility on its part.

After Medic, Danny Sullivan from Google directed webmasters towards these guidelines. His message has since disappeared, but Marie Haynes recounted the advice at the time. The founder of the firm claims to have spent months studying these recommendations and shared at BrightonSEO in 2019 that she keeps a printed and laminated copy on her desk.

Algorythmic Bases Its Method on Experts

The founder of Algorythmic advocates a similar thesis: identifiable experts and their experience provide a more sustainable response to generic AI content than processes aimed at exploiting search engines. This approach, recommended for several years, has been named the "E-E-A-T Method" and forms the philosophy of the new firm. It comes with a warning against mass publication for search engines without human contribution.

According to the founder of the firm, this approach requires the participation of the company's experts, their time, and resources, rather than limiting work to the SEO team. She attributes benefits to users, conversions, and other channels, and believes it also seems to help with AI systems. She considers the risk of penalization to be low, provided that this approach is not applied with a shortcut mentality.

This same individual dedicated an article in 2021 to the misinterpretations of E-E-A-T. At MozCon 2022, her presentation focused on real expertise, with examples of specialists representing the brands for which they worked. She presents this framework as the most suitable for reducing exposure to devaluation or subsequent penalization.

The Risks of Shortcuts in SEO

The practices targeted by these warnings cover several generations of SEO: exact match domains, article directories, parasitic SEO, massive guest posting, private blog networks, and programmatic pages. This also includes AI-generated lists, large-scale automated production, and inauthentic mentions aimed at visibility in AI responses. The founder of Algorythmic believes their effectiveness can disappear when search engines or AI companies intervene, leading to prolonged losses in traffic, visibility, and business contacts.

This specialist also reports receiving companies in difficulty after massive AI translations with little proofreading. She compares the methods of certain content production tools to practices previously targeted by Google, asserting that several sites used in their commercial examples quickly lost organic traffic.

She also shares a personal precedent: gains attributed to link buying in 2010-2011 were wiped out by Penguin in a single morning in 2012.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.