Google's March 2024 Core Algorithm Update penalizes sites created with AI-generated content, but AI-generated content does not meet quality standards outlined in various Google documents. It's a fact. But there are still ways to use AI to produce high-quality content.
Why AI can't meet Google's quality standards
Some ranking systems, including reviews and helpful content systems, explicitly set quality standards that are essentially impossible for AI-generated content to meet.
Adding an extra E (for experience) to the EAT should have sent a message to content creators that using AI comes with risks.
Examples of SERP features, quality signals, and ranking signals that inherently exclude AI content
The writing on the wall regarding AI content has always been visible.
Here are some qualities that Google's documentation says are important to filter out content that is purely AI-generated:
- experience
- Published reviews must be practical
- Google News prioritizes human authors in Google News search results
- Announced in May 2023, Google Perspectives highlights human authors (hidden gems) found in forums
- Author page (specialized questions)
- Author background information (professional questions)
- Author introduction page (professional questions)
quality concept
Google has published self-assessment questions to help publishers see if their content meets Google's quality standards.
These questions do not have any specific ranking factors listed. They only enumerate concepts that reflect the trends that high-quality websites generally exhibit.
If AI-generated content does not fit these concepts, it may not meet quality standards, even if publishers try to disguise outward signs of quality, such as author pages.
Author and expertise
The expertise section of a self-assessment document mentions the author in a way that machine-generated content cannot reproduce.
This section states:
“Content presents information in a way that makes people want to trust it, including clear sources, evidence of relevant expertise, and context about the author and the site that publishes the content (such as an author page or a link to the site). What about the page?”
The sections cited above focus on expertise in three elements:
- Sourcing (citing sources, fact checking, attribution of quotations)
- Evidence of expertise involved
- Author's biography
These three qualities are outward signs typically associated with expertise that cannot be achieved with AI.
Content quality: originality
The content and quality sections of your self-assessment guide should be unique.
That section of Google's documentation asks for the following:
“Does the content provide original information, reporting, research, or analysis?
…Does your content provide insightful analysis or interesting information that goes beyond the obvious?
Lack of originality is a hallmark of generative AI. The content created by generative AI is literally the most likely set of words regarding a particular topic.
first-hand expertise
The people-first section of the self-assessment questions asks about first-hand expertise.
“Does your content clearly demonstrate first-hand expertise and deep knowledge (e.g., expertise gained from actually using the product or service or visiting the location)?”
Obviously, the machine has no direct expertise. You may not handle the products or use the services.
AI can still be used for content creation
Given that many sites with AI-generated content received manual action during the March 2024 core algorithm update, it may be time to reconsider the place of AI in web content. maybe.
There are still ways to leverage AI to produce high-quality human-first content. The most important thing about content is not who wrote it, but the insight behind it.
The way forward might be to combine human insight and experience and use it as data that AI can use to generate content.
How to create review content using AI
For example, you can expand your product reviews by creating a checklist of data points that consumers need to make a purchasing decision. Someone needs to handle and review the product, but the reviewer only needs to write a score and comment for each data point on his checklist.
If the review is for a children's bike, benchmark what people want to know about the bike, such as the ages and sizes the bike is suitable for, how much it weighs, and how sturdy the training wheels are. For TV reviews, the checklist includes benchmarks for things like richness of black levels, off-center display, and ease of setting colors.
At the end of the checklist is a section called Final Impressions, which includes the pros and cons as well as whether the reviewer has a positive, neutral, or negative opinion about the product, and overall information about who they feel the product is best for. List your emotions. People looking for budget, performance, etc. Once that's done, upload your document to the AI and ask it to write a review.
How to write any kind of content using AI
A friend of mine gave me a tip about using AI to refine rough content. His workflow consists of dictating everything he needs to say into a recording, regardless of paragraph structure, and simply pouring it into the recording. Then upload it to ChatGPT and ask them to turn it into a professional document. Additionally, you can ask them to create pros and cons and summaries.
AI amplifies human input
My suggestion is to think of AI as a ghostwriter that turns rough documents into polished essays and articles. This approach works in almost any scenario, including scaled product descriptions.
There are important qualities of content that humans can provide that AI cannot, such as sources of information, evidence of expertise, sources, and the context humans bring to the topic being written about. Humans bring experience, expertise, authority, and credibility. AI can take elements provided by humans and turn them into high-quality content.
Given that many sites with AI-generated content received manual action during the March 2024 core algorithm update, it’s time to rethink how AI is used with respect to content. It may be that you are coming to.
I planned and wrote most of this article in September 2023, but I left it alone because I wondered who would believe me.
Now in March 2024, people are becoming more proactive in considering better ways to integrate AI into their content generation workflows, as the SEO industry faces calculations based in part on AI-generated content. It may be.
Featured image by Shutterstock/Roman Samborskyi