Google doesn’t have a problem with AI-generated content. What it does have a problem with is empty content – content that lacks substance, originality or usefulness. Google’s goal is to limit empty content.
As Google quality raters are now tasked with evaluating AI-generated content, it might be a good time to look at what the search giant considers to be “bad content”.
What Google really wants
Despite popular assumptions, Google has no interest in making its search service AI-free. The tech company has nothing at all against AI in principle. If it did, it would do whatever was in its power to sink the boat, rather than vigorously wanting to jump on it, like everyone else these days.
In reality, Google is interested in distinguishing high-value, intellectually rich content from filler, or empty, content.
So, the real question we should be asking is: Can AI create good content?
Where AI falls short in content creation
Personally, I don’t think so. There is a very precise divide between the huge mountain of junk content online and the minimal amount of valuable, original content. This divide can be summed up in a single word: opinion.
Large Language Models (LLMs), like Gemini or ChatGPT, don’t actually create—they copy. We can debate whether it will be able to do so in the future, and I humbly doubt that as well, but it is clear that they are currently not capable.
At the moment, what LLMs are able to do is try to predict what a non-AI source might write in a given context and in many instances, it does so beautifully, almost magically.
But alas, they don’t generate anything new. They don’t express ideas. They lack opinions because they lack consciousness.
Take, for example, this very article that you’re reading: it is precisely what any AI cannot create. I am presenting an idea that I had, one upon which I wanted to develop.
Since I have not read it in many places I thought of writing it down and publishing it on an authoritative site, in order to expose my idea to my peers and spark discourse.
Whether you agree or disagree is irrelevant. It doesn’t matter at all how popular my opinion is. It doesn’t even matter if it is actually original or not. (Unbeknownst to me, someone may have already written about it before and in fact it is very likely, since the internet is very big.)
What matters is that at its core, my objective is to express an idea that originated from personal insight – good or bad. What matters is that I am expressing an opinion.
AI cannot have an opinion, that’s something it can’t replicate.
By design, it is great at copying what humans do, the intellectual process of absorbing information from external sources and sharing existing knowledge.
However, it cannot do that one thing that sets humans apart, it cannot generate a new angle or formulate an opinion of its own. This limitation is foundational – it’s tied to the hard problem of consciousness, a topic beyond the scope of this article, but important for the point I’m trying to make.
[Case Study] How GroupM Turkey helped IKEA increase their non-branded organic traffic
Google’s contradictory position on AI
There are, of course, strong economic reasons behind Google’s need to maintain a dichotomous approach to the use of AI for content production: they must both champion AI and restrict it. I’ll focus on textual content here, but what I write applies to any type of content.
On the one hand, they need to promote it, push it on the user with such stubbornness and invasiveness that it seems like the idea of wanting to use AI for anything comes from the user themselves, and not from Google et similia. First comes the investment, then must come the user’s need.
On the other hand, they more or less need to prevent AI content from flooding the SERPs.
Here, I’ll go on a brief tangent about the ‘more or less’. Google no longer needs its search engine to be the best, since many users now default to Google out of habit. Therefore, the SERPs not being at the highest possible qualitative level may no longer be necessary.
Whether this is a short-sighted policy is up for debate (it is), but at the moment Google doesn’t seem to care too much, or at all.
Yes, they want to keep the public on the search engine, but the quality of the results is less and less important to them, as I think is clear to anyone who has done any research in recent times.
This line of reasoning comes from witnessing the SERP’s transformation in time. It has become increasingly similar to the competitors that failed against Google in the last 20 years: less minimalist, more flashy, full of distracting features and paid areas. Less helpful for the users, more profitable for the paying brands.
All that being said, Google’s aim is to prevent a surplus of low-quality content that could erode trust and further degrade the user experience— the very thing Google’s algorithms are designed to protect.
The double standard
Anyway, back to the central point: there is a contradiction in Google’s policy. The search engine can use AI to share search, but a webmaster shouldn’t use it to shape content. Why?
To understand the motivation that pushes and forces Google into this ambiguous position, we need to think about the nature of the textual content generated by AI, and the nature of AI itself.
Let’s start from the premise that humans are not simply interested in reading text. A human, rather, likes/is interested/finds use in absorbing information.
Information can reach us from external sources through our senses and reading is a great way to absorb information – so good that it’s been decisive for the development of the species itself.
Seeking information is the reason why we use the internet, as well as Google. Before AI, the internet was already full of useless and redundant content. Those of us who have been online for years know all too well: empty content was not born with AI, it has always been there. Remember ‘manual’ content spinners? Aren’t they just AI 0.1?
I think it’s fair to say that the spam produced via spinning ten years ago has the same value for Google and the user as the AI-generated content of today: none.
Google has always tried, and will continue, to limit poor content with its spam policies.
If this article is sufficiently original, in the sense that I am not repeating something that’s already been said a thousand times, then I deserve to bring traffic to the Oncrawl blog.
If instead I am just putting words in a row without much added value, then I do not deserve attention. There are a million self-appointed “SEO experts” who publish articles on AI.
It would be akin to anyone trying to rewrite Shakespeare: Shakespeare deserves attention, those who reproduce him do not.
The fundamental truth
Now, this is the fundamental point: AI can copy Hamlet, but it can’t create it. Ever.
What content creators need to understand
If there is one takeaway for SEOs and content writers, it’s this: in order for their online content to be successful, what they publish needs to say something useful to those who read it.
This has always been the case and it will continue to be so for the foreseeable future.
Is it possible to provide the audience with interesting content using AI? Of course! If you start with a strong idea, AI can help structure, polish, and expand your ideas. However, keep in mind that quality and usefulness are what matter most with any content we produce.
Final thought
AI content is not the enemy – bad content is.
We need to prioritize quality, originality, and usefulness. Remember to write for people, not algorithms and you should be fine!
