Google’s algorithm hates AI content? How to like Google?


Google’s March 2024 core algorithm update is penalizing sites made of AI generated content and it is a fact that AI-generated content may not meet the quality standards that are mentioned in various Google documents. But there is still a way to use AI in a way that produces high-quality content.

Why can’t AI meet Google’s quality threshold?

Many ranking systems, including review and supporting content systems, apparently have quality standards that naturally make content written by AI impossible to satisfy.

The inclusion of an extra E in EAT (for experience) should have been a signal to content creators that using AI carries risks.

Examples of SERP features, quality signals, and ranking signals that naturally populate AI content

The writing on the wall regarding AI content has always been in clear sight.

Here are some of the properties that Google’s documentation says are important in ruling out AI-generated content entirely:

  • Experience
  • Published reviews must be practical
  • Google News emphasizes human writers in Google News SERPs
  • Google Perspectives, announced in May 2023, emphasizes human authors (hidden gems) found in forums
  • Author Page (Expertise Question)
  • Author Background Information (Expertise Questions)
  • About Author Page (Expertise Questions)

quality concepts

Google published self-assessment questions to help publishers identify whether their content conforms to Google’s quality standards.

These questions do not list specific ranking factors. They simply list concepts of things that generally reflect what high quality websites show.

If AI-generated content cannot fit into those concepts then it is possible that the content does not meet quality standards, even if publishers try to fake external signals of quality such as author pages, etc.

Authorship and expertise

The expertise section of the self-assessment document cites authors in a way that cannot be replicated by machine-generated content.

This section states:

“Does the content present information in a way that makes you want to rely on it, such as clear sources, evidence of expertise involved, background about the author or the site that published it, such as a link to an author page or a site Via “About” page?

The section quoted above focuses expertise on the following three factors:

  1. Sourcing (citation of sources, fact checking, attribution of quotes)
  2. Evidence of expertise involved
  3. author background

Those three qualities are external signals typically associated with expertise that cannot be obtained by AI.

Content Quality: Originality

The content and quality section of the self-assessment guide requires originality.

Here’s what that section of Google’s documentation asks:

“Does the content provide original information, reporting, research or analysis?
…Does the content provide insightful analysis or interesting information that goes beyond the obvious?”

Originality is the hallmark of generative AI. Content created by generative AI is the most likely string of words on virtually any topic.

first-hand expertise

The mass-first section of self-assessment questions asks about direct expertise:

“Does your content clearly demonstrate direct expertise and depth of knowledge (for example, expertise that comes from actually using a product or service or visiting a location)?”

Obviously no machine has direct expertise. It cannot handle any product or use any service.

AI can still be used for content creation

Given how many sites with AI-generated content are taking manual action during the March 2024 core algorithm update, it may be time to rethink the place of AI for web content.

There is still a way to use AI that can result in high quality people-first content. What matters most about the content is the insight behind the content, not who or what wrote it.

One way forward could be to use a mix of human insight and experience as data that AI can use to generate content.

How to Create Review Content with AI

For example, it is possible to measure product reviews by creating a checklist of data points that consumers need to make a purchasing decision. Someone will still have to handle the product and review it, but they will have to write scores and comments for each data point on the review checklist.

If the review is of a children’s bicycle, benchmark things users would want to know about the bicycle such as what ages and sizes the bicycle fits, how much it weighs, how strong are the training wheels, etc. If this is a television review the checklist will contain benchmarks related to richness of black levels, off-center viewing, ease of setting colors, etc.

At the end of the checklist there is a section called Final Impressions which lists the pros and cons as well as the overall feeling where reviewers write whether they feel positive, neutral, negative about the product and whether they think it The product is best for people like them. A budget, those who crave performance and so on. Once this is complete, upload the document to your AI and ask it to write a review.

How to Write Any Type of Content with AI

An acquaintance shared a tip with me about using AI to improve rough content. His workflow involves dictating everything that is said in a recording, regardless of paragraph structure, and then simply putting it into the recording. Then he uploads it to ChatGPT and asks to convert it into a professional document. He may also ask for its pros and cons and to prepare an executive summary.

AI augments human input

I suggest thinking of the AI ​​as a ghostwriter that takes a raw document and turns it into a polished essay or article. This approach can work for almost any scenario, including scaled product descriptions.

The important qualities of content are those provided by a human that AI is unable to do, things like sourcing, evidence of expertise, sourcing, and the background that a human brings to the topic being written about. Humans bring experience, expertise, empowerment and trustworthiness. AI can take those elements provided by humans and turn it into high-quality content.

Given how many sites with AI-generated content are being slapped with manual tasks during the March 2024 core algorithm update, it may be time to rethink how AI is used when it comes to content. is done.

I planned and wrote most of this article in September 2023 and sat on it because I thought, who will believe me?

Now that it’s March 2024 and the SEO industry is facing a reckoning partly based on AI-generated content, people may be more receptive to considering better ways to integrate AI into the content creation workflow.

Leave a Comment