The State of Google SEO & AI Today: What Do You Think?

8 replies
Originality.AI published this report: https://originality.ai/can-google-de...ize-ai-content
And, here's what I think, based on how they wrote that report and how they designed their two research studies:

Well I don't know about you, but when it comes to Google's SERP algos -- I for one tend to trust what they themselves publicly announce and consistently publish.
Meanwhile, I tend to have a healthy suspicion on groups that clearly have a vested commercial interest in contradicting what Google's saying.

More so if that group publishes a poorly executed research study that doesn't factor in what the algo designers themselves say they're consistently implementing over the years.
For instance, A group with a clear commercial interest in having more content publishers believe they should check their AI-generated content for "human" or "AI" percentages shouldn't design nor publish a research study that focuses solely on AI-generatedness as an end-all-be-all correlation metric for de-indexing.
That's because the same group should've focused more on spamminess as a metric with more weight, then analyzing results to scientifically find out if any measurable correlation exists between AI-generatedness and de-indexing.
Especially if the algo designers themselves stated that spamminess, helpfulness, timeliness and relevant value for readers have always been their ranking metrics.

Supplemental Info
Here's what Google publicly announced last year (Feb 8, 2023), which they have NEVER recanted as of today:

QUOTE:

At Google, we've long believed in the power of AI to transform the ability to deliver helpful information.
In this post, we'll share more about how AI-generated content fits into our long-standing approach to show helpful content to people on Search.

Rewarding high-quality content, however it is produced
Our focus on the quality of content, rather than how content is produced, is a useful guide that has helped us deliver reliable, high quality results to users for years.
For example, about 10 years ago, there were understandable concerns about a rise in mass-produced yet human-generated content.
No one would have thought it reasonable for us to declare a ban on all human-generated content in response.
Instead, it made more sense to improve our systems to reward quality content, as we did.

Focusing on rewarding quality content has been core to Google since we began.
It continues today, including through our ranking systems designed to surface reliable information and our helpful content system.
The helpful content system was introduced last year to better ensure those searching get content created primarily for people, rather than for search ranking purposes.

How automation can create helpful content
When it comes to automatically generated content, our guidance has been consistent for years.
Using automation--including AI--to generate content with the primary purpose of manipulating ranking in search results is a violation of our spam policies.
Google has many years of experience dealing with automation being used in an attempt to game search results.
Our spam-fighting efforts-- including our SpamBrain system -- will continue, however spam is produced.

This said, it's important to recognize that NOT ALL use of automation, including AI generation, is spam.
Automation has long been used to generate helpful content, such as sports scores, weather forecasts, and transcripts.

AI has the ability to power new levels of expression and creativity, and to serve as a critical tool to help people create great content for the web.
This is in line with how we've always thought about empowering people with new technologies.
We'll continue taking this responsible approach, while also maintaining a high bar for information quality and the overall helpfulness of content on Search.

Our advice for creators considering AI-generation
As explained, however content is produced, those seeking success in Google Search should be looking to produce original, high-quality, people-first content demonstrating qualities E-E-A-T.
Creators can learn more about the concept of E-E-A-T on our Creating helpful, reliable, people-first content help page.
In addition, we've updated that page with some guidance about thinking in terms of Who, How, and Why in relation to how content is produced.
Evaluating your content in this way, whether you're using AI-generated content or not, will help you stay on course with what our systems seek to reward.

Posted by Danny Sullivan and Chris Nelson, on behalf of the Google Search Quality team

/QUOTE
source = https://developers.google.com/search...nerate-content

Additional Data
Here's a report with qualitative data that was published by SEMRush on Jan 31 this ear ...

QUOTE:

The study analyzed over 2,600 businesses globally providing critical insights into the current state and future trends of AI in content marketing.
67% of businesses already use AI for content marketing and SEO, and 78% are satisfied with the results.

Key Takeaways
** 65% of businesses generate better SEO results thanks to AI.
** 67% also see an improvement in content quality when using AI.
** 68% get a higher content marketing ROI due to AI.
** 37% of companies that don't use AI don't understand how it works.
** 93% review their AI-generated content before publishing it.
** The majority of consumers tend to prefer AI-generated copy.

In other words, businesses using AI tools report higher effectiveness in their content marketing efforts and better results in attracting organic traffic.

/QUOTE
source = https://www.semrush.com/news/288870-...ends-for-2024/

So, what do you think?
#AI Content #google #google seo #originality.ai report #semrush report #seo #seo ai #state #today
Avatar of Unregistered
  • Profile picture of the author savidge4
    I dont want anyone to get me wrong... AI has its place - even in content creation.

    The thing is pick a topic any topic and have AI re-write the same content a number of times over a number of days or weeks and the content ends up being somewhat the same - and the question begged to be asked is who would write the same article say 10 times? The answer is any one person is not going to do this, but 10's of thousands of individual people will write the article once individually.

    AI is only as good as the data it is given... granted in content creation we are in essence suggesting there is an entire internet of content... but is this realistically a reality? Pick an item, any item on say Amazon, have AI write a description for said item... will the AI description be anything different than what is already there? Technically it cant, right?

    AI in regard to content creation is in a way dumbing down the level of content available < ducking> it can ONLY rehash and reproduce what is already there ( its given data source )

    Google says hey we dont want SEO Spam content and yeah you can use AI to produce that type of content... But you have to remember Google also says it is looking for Unique content or maybe a term they use would be "High Quality" content.

    If you are writing about something you dont know... yeah sure AI is a no brainer... the moment you fall into a subject that you have either Knowledge or Experience I think AI might be a starting place in the creation process... but it by no means should be the end all definitive

    Adding Conjecture or an Opinion on a subject you have Knowledge or Experience about... this content becomes unique - and probably by comparison to the standard AI results, a better more informative read.

    Again dont get me wrong, the cost for entry to AI vs say having non English native people writing your content for pennies is worth the dollars spent with AI... yeah, you are going to get better content - and that means ( ish ) you will get better results. Like I suggested earlier I believe this will dilute the content across the board as a whole, and those with actual Knowledge and Experience communicating on a subject will have the ability to provide a less vanilla product - which should result in a better end user experience.

    Those of us that know AI... understand a closed set of data is far more accurate than the jumbled mess that is the internet. "Internet Data" is absolutely flawed on a whole lot of levels. learn and understand "link rot" and you then understand why College students writing papers have bad notations - AI replicating what is within its data set - broken links - Again AI is only as good as the data it is provided - and it is duplicated the known pattern of behavior that is Link Rot.

    AI will improve the readability of the internet as a whole - without question - but as it stands right now and into a not so near future, it will not produce ground breaking unique content with new thoughts and ideas - it cant
    Signature
    Success is an ACT not an idea
    {{ DiscussionBoard.errors[11789813].message }}
    • @savidge4,

      Originally Posted by savidge4 View Post

      The thing is pick a topic any topic and have AI re-write the same content a number of times over a number of days or weeks and the content ends up being somewhat the same - and the question begged to be asked is who would write the same article say 10 times? The answer is any one person is not going to do this, but 10's of thousands of individual people will write the article once individually.
      When I was still employed in 2004 as a content writer at a local BPO here in Manila, I was asked to write SEO content about blackjack, something that I have specialist experience in, not to mention formidable authority knowledge that's relevant for the game (my undergrad thesis in mathematics involved probabilistic statistics and game theory).

      So no problem there.

      But the rub was, I was given the same set of keywords in the same exact niche topic and was asked to write 4K++ words of content each day, for 3 months straight.

      So yes, I agree.
      AI. Human. Hybrid AI-human.
      It doesn't matter.
      Nobody should do this.
      Though many if not most SEOers had this idea then.
      and many still do, up 'til today.
      Plus, many site and business owners still have the same exact idea when it comes to SEO.
      I should know. I continue to run a business serving a lot of them in North America, the EU and ASEANA for the past two decades.

      Originally Posted by savidge4 View Post

      AI is only as good as the data it is given... granted in content creation we are in essence suggesting there is an entire internet of content... but is this realistically a reality? Pick an item, any item on say Amazon, have AI write a description for said item... will the AI description be anything different than what is already there? Technically it cant, right?
      But neither could many humans.
      Even with authority knowledge and specialist expertise.

      And the thing here is, many use the same exact AI model for very similar SEO content generation use cases.
      they use it unchanged, untouched.
      They don't cherrypick, test and optimize open source, proprietary and custom models (custom = use foundational models to develop your own via zero-shot knowledge distillation).
      They don't develop, customize nor even fine-tune it to have authority knowledge and specialist expertise specific to their niche and use case.
      They don't create a programmatic workflow that simulates human creative workflow also specific to their marketing communication strategy for their particular target geos and niche audiences.

      For example, let's say you create and customize this programmatic workflow:

      Automatically harvest then input a strategically formulated query (strategic = since you know how your model performs because you developed, tested, customized and fine-tuned it yourself for your specific use case) with the target exact match keywords, longtail keywords, question-type keywords, metadata of the top X ranking pages in Google for those terms and the target geos, thoughtfully crafted details about relevant stuff like voice, ideal audience avatar details and so on;

      >

      Automatically harvest research data sources relevant to the current inputs from Google or more so from other content repos that serve relevant, timely, factual and technically accurate content specific to your target niche, geos, ideal viewers and so on, i.e. Peer-reviewed academic journals and repos of published research papers specializing in your target subject areas;

      >

      Process those inputs the way you want to through strategic prompts, with one in a set of objectives being to outdo the inputs (like the metadata of the top ranking pages, for example), another one being to strictly enforce your formatting and outlining instructions, another to mimick the voice and style that you want (which your AI model should've been fine-tuned on from the get-go), and the primary goal being to generate content that complies with Google's E-A-T quality standards (which you should've fine-tuned your AI model on from the get-go); and

      >

      Activate the last phase of your pipeline, which is for a human with authority knowledge, specialist expertise, direct response copywriting skills, content editing, design and CMS formatting experience for SE-optimization to check, edit and post the AI-generated content.

      Well, the benefits of that hybrid AI-human pipeline that my clients and I have been experiencing from Dec 2022 up until today outweigh the positive stuff that we've been getting from a purely human pipeline.

      So our results are similar to the 78% of the 2K++ businesses that participated in SEMRush's report.

      And upon reading that pipeline again, that's a very, very similar creative human-only process for achieving the target objectives with considerable success rates.

      But yeah, I think lots of lots still don't know how to work this technology to get benefits for their particular use cases better than not using it at all.

      And there's the rub. They'll keep doing what they're doing now. Mainly because their idea is to mass produce content to hit top Google spots.

      Keep on throwing stuff, and see what sticks, they'd tell you, without even considering that the stuff they're throwing is likely mud, or something that stinks more.

      Anyway, my point is, we should all really sit down and think of what to do so as to best benefit from any tech.

      The solution of sighted people back then for blind persons to achieve the highest possible mobility independence is to use a human guide and a wheelchair.

      That was until someone gave a blind person a stick and cardboard with holes that can be felt with fingertips. And the rest is history.

      Cheers!
      Signature
      Need Custom Programmatic SEO or GenAI Engineering Work Done? Drop Me an Email HERE ...
      • Chief Machine Learning Engineer @ ARIA Research (Sydney, AU)
      • Lead GenAI SEO Campaign Engineer @ Kiteworks, Inc. (SF, US)
      {{ DiscussionBoard.errors[11789819].message }}
      • Profile picture of the author savidge4
        Originally Posted by Marx Vergel Melencio View Post

        So our results are similar to the 78% of the 2K++ businesses that participated in SEMRush's report.
        Lets start here... I have to just wonder, how many of the 2000+ business' are using a custom programmatic workflow? 1 in 10? 1 in 20? 1 in 100? it sure isnt 100%, and I would venture to say its less than 10%. Fixing linguistic errors in un itself probably could account for the uptick.

        I have said for years on this very forum that SEO really isnt that hard... once you apply an amount of data.

        You and I in a whole lot of ways are saying the same thing... Picking and choosing data sets... in essence "closed data sets" vs the internet at large will only increase a piece of contents ability to #1 communicate on the Knowledge and Experience level, #2 separate that piece of content from the sea of re-spit internet mush, and #3 increase the reading level of that piece of content.

        En masse you end up with a vast sea of have nots, vs the way smaller percentage of people that have the technical ability that get vetted, tested and closed data sets that have the ability to create good content.

        The line that separates the 2 ( haves and have nots ) is the ability to program.

        I will more than assume your content passes the "Turnitin" test.... meaning its not detected. That obviously doesn't happen by accident.

        So no, people dont understand the "Power" of AI... what they understand is the ease of use - BUT, easy is not Better. BUT, easy is better than what they were producing prior to the likes of Chatgpt. And this is why a guy like you is taking it to the bank - because you are bypassing the pitfalls of AI "out of the box"

        I personally use AI with my web business to analyze data and react accordingly. Something as mundane as determining based on MAC how many times a person visits a site prior to purchase, to say analyzing visitor IP addresses to determine geographically where to target say Facebook ads - your basic in house self modeled closed data type stuff. we do an amount of AI generated content using a very similar programmatic system as you have laid out, more often than not it is quicker and "easier" to just write the content as needed, because I have basically 1000+ local based small business that do 1000 different things. We focus more on the conversion side of things, and YES, use AI to parse that data. Traffic isnt my greatest issue ( I got 99 problems and traffic aint one ) Converting is where focus the greatest amont of time

        AI is a tool - and maybe i can take that a step further... AI is a toy, and it takes the ability to "prompt it" - I say more program it - to utilize it as a tool.
        Signature
        Success is an ACT not an idea
        {{ DiscussionBoard.errors[11789824].message }}
        • @savidge4,

          Originally Posted by savidge4 View Post

          [SNIP]
          Fixing linguistic errors in un itself probably could account for the uptick.
          [/SNIP]
          Yes, I agree. My wager's a majority of those 2K++ businesses is getting more value mainly (and probably only) because they're now able to hire one much better content editor with authority knowledge, specialist expertise and AI whispering experience at $80K to replace 3 or so generalist content writers at $36K each to produce the same content volume at much higher quality for their SEO targets.

          Originally Posted by savidge4 View Post

          [SNIP]
          You and I in a whole lot of ways are saying the same thing... Picking and choosing data sets... in essence "closed data sets" vs the internet at large will only increase a piece of contents ability to #1 communicate on the Knowledge and Experience level, #2 separate that piece of content from the sea of re-spit internet mush, and #3 increase the reading level of that piece of content.
          [/SNIP]
          Yes. And I can't understand why many businesses fail to think about this from that perspective.
          I mean, a specialist in anything tends to always be the better choice over a generalist, across the board.
          And that's even if all 'em same business owners will say that they'll consider the advice of a veteran surgeon who has been specializing over the past few decades in the particular operation that their loved one needs, over advice about the same thing from a general physician interning for just the past couple of months.

          Originally Posted by savidge4 View Post

          [SNIP]
          So no, people dont understand the "Power" of AI... what they understand is the ease of use - BUT, easy is not Better. BUT, easy is better than what they were producing prior to the likes of Chatgpt.
          [/SNIP]
          The hype about this tech is also to blame from how that kind of thinking continues to thrive.

          Too many businesses and generalist content writers believe they can just ask ChatGPT to write a 2K++-word content post about a niche topic they know nothing about with just one conversation turn (one prompt).
          More so, they expect the post to be SE-optimized, have design elements to improve visual impact like tables and suitable images, have an easily skimmable header outline that strategically divides the content into sections designed with maximum effect from the implementation of suitable copywriting frameworks, and so on.

          And that thinking isn't exclusive to just the SEO content writing space.
          For example, many with little to no fundamental knowledge on Python and C++ are expecting ChatGPT to be able to, with just one conversation turn, convert Python code that uses multiple third party libraries, sync and async functions that place API calls to different Cloud providers into a C++ binary.
          If the binary that ChatGPT generates doesn't work, which almost surely won't, they'll say ChatGPT doesn't understand Python and C++.
          They don't think about asking themselves the obvious (how do they know this for sure since they don't have significant experience in these languages).
          ** I've been testing similar tasks, and it takes me an average of 15 conversation turns to do this correctly with ChatGPT, and I have significant experience in Python and C++ already.

          As for CRO, I think AI's utility for this use case, aside from analyzing relevant data points as its primary use which you mentioned, is to provide the human copywriter with ideas that could help.
          And not the other way around -- It's not feasible at the moment to rely primarily on AI with human as secondary support to do this.
          Signature
          Need Custom Programmatic SEO or GenAI Engineering Work Done? Drop Me an Email HERE ...
          • Chief Machine Learning Engineer @ ARIA Research (Sydney, AU)
          • Lead GenAI SEO Campaign Engineer @ Kiteworks, Inc. (SF, US)
          {{ DiscussionBoard.errors[11789835].message }}
          • Profile picture of the author socialentry
            Originally Posted by Marx Vergel Melencio View Post

            They don't think about asking themselves the obvious (how do they know this for sure since they don't have significant experience in these languages).
            if matrices can understand C++ do they dream of electric sheeps
            {{ DiscussionBoard.errors[11789944].message }}
            • @socialentry,

              Hmmm ... Only if they've overloaded their constructor to include imagination.
              But instead of electric sheep, they probably dream of arrays of electric goats.
              After all, it's all about memory optimization ...

              Though thinking about this more, I'd say if matrices can profoundly understand C++, then from a probabilistic statistics standpoint, their dreams of electric sheep could be modeled as a Markov Chain.
              That's where the state of each dream sequence is dependent only on the previous state.
              And essentially, matrices dreaming of electric sheep would be an iterative process, navigating through a stochastic matrix of dream scenarios.
              That's where each electric sheep encounter is a transition towards achieving an optimal dreaming state, underpinned by Eigenvalues of relaxation and Eigenvectors, pointing towards the direction of the most surreal dreams ...

              C'mon now. Everybody knows this!
              Signature
              Need Custom Programmatic SEO or GenAI Engineering Work Done? Drop Me an Email HERE ...
              • Chief Machine Learning Engineer @ ARIA Research (Sydney, AU)
              • Lead GenAI SEO Campaign Engineer @ Kiteworks, Inc. (SF, US)
              {{ DiscussionBoard.errors[11789952].message }}
  • Profile picture of the author socialentry
    It failed. Some of it is simply rot, walled gardens and out of date info.
    But for them most part ,humans have gamed the system . They produced garbage in mass thats acceptable to the G. AI is not going to move the needle.



    infity garbage times 2 times infinity is still infinity


    Its even worse then 20 years ago when there were obvious link farms because now you have to sift through stuff that looks plausible.

    You just get beginner content over and over again.

    Its just mountains and mountains of over-copywritten garbage. Endless streams of "10 ways to do XYZ" with pretty graphics and websites that demand 4 GB of RAM so they can tell you tommorow is going to be rainy.

    The optimal G query is "your query reddit".
    {{ DiscussionBoard.errors[11789814].message }}
  • Profile picture of the author Monetize
    Originally Posted by Marx Vergel Melencio View Post

    Key Takeaways
    ** 65% of businesses generate better SEO results thanks to AI.
    ** 67% also see an improvement in content quality when using AI.
    ** 68% get a higher content marketing ROI due to AI.
    ** 37% of companies that don't use AI don't understand how it works.
    ** 93% review their AI-generated content before publishing it.
    ** The majority of consumers tend to prefer AI-generated copy.

    In other words, businesses using AI tools report higher effectiveness in their content marketing efforts and better results in attracting organic traffic.

    So, what do you think?

    I think that your post is very informative, so thank you for
    taking the time to research and develop it.

    Regarding AI, I have heard that Google doesn't care and
    that it only requires that content is helpful, and your post
    confirms that theory.

    For me, AI makes me more productive than I ever thought
    I could be and it helps me create all sorts of content and
    products, that I never would have been able to do.

    Even if I hired content creators, as I have done in the past,
    you never know what you'll receive or where their content
    came from. I could write a whole book about my nightmare
    experiences with people that I hired before.

    I also love being able to brainstorm and strategize methods
    anytime I want, and how AI helps me organize things like
    schedules, projects, as well as giving me step-by-step lists
    and instructions for getting things done.

    Humans moaning about its limitations probably don't know
    how to optimize it, it helps if the operator can think outside
    the box and provide AI with the necessary guidance to get
    the results they need.

    Anyway, thanks again for this fantastic thread.
    {{ DiscussionBoard.errors[11789822].message }}
Avatar of Unregistered

Trending Topics