Between millions of automatically generated articles and a network increasingly emptied of human presence, is the Internet really dying? Will the web become an empty shell populated only by bots?
A new episode of SEO Confidential, the space where the most authoritative voices in search marketing (and beyond) speak without filters, bringing perspectives that, even when they don’t always match our own, allow us to broaden our horizons, step outside our certainties and understand where online search is really heading.
This series was created for exactly that: to bring together the ideas, experiences and intuitions of those who live SEO and the evolution of artificial intelligence on the front line, day after day.
The guest of this episode is Myriam Jessier, an international consultant with over fifteen years of experience in search engine optimization and performance marketing.
She has worked with brands across very different industries – from finance to fashion, from food to public administration – and her name has appeared in publications such as Semrush, Oncrawl and Smashing Magazine, as well as on the stages of conferences like BrightonSEO, OMR, Voxxed Days and SMX.
Today Myriam supports global companies in defining strategies that combine data, technical branding and artificial intelligence, turning the complexity of algorithms into operational decisions capable of generating value.
In this interview we tackle topics that are reshaping the present and future of search: the increasingly intrusive role of AI in optimization and content creation, the risks of LLM-driven browsers, the phenomenon of blanding and the aesthetic standardization that is flattening the web, all the way to the legal tensions running through platforms, creators and AI companies.
A lucid, provocative and necessary journey into the contradictions of search in the AI era.
“We are converging into one big pile of bland, cliché-ridden aesthetics. Only human beings can create narratives that spark curiosity”
The number of articles written by artificial intelligence has surpassed those written by humans, but their growth has stalled and most of these texts never appear on Google or ChatGPT. So what is the point of producing so much automated content if nobody sees or reads it?
What if this is web 3.0? A bloated, dying corpse with a corpus overwhelmingly generated by artificial intelligence?
The Dead Internet theory is an online conspiracy theory claiming that the Internet now consists mainly of bot activity and automatically generated content manipulated by algorithmic curation, marginalizing organic human activity.
In a way, that is true. And once the Internet dies, it becomes much easier to destroy human communication at scale. It is easy to break grassroots movements if they have no place to organize. I am worried about the future.
The so-called “AI bubble”, inflated by billions in investments and promises of AGI, looks more and more like the dot-com rush: how long before the market stops funding illusions and starts demanding concrete results?
I think we could be 2-5 years away from a market correction when it comes to the AI bubble. It reminds me of the Soviet Union. It will keep going without admitting any flaw or failure until it collapses.
What will remain, however, is the solid part of this technology, the part you can truly rely on to build new ways of creating and developing services on the web.
In a web where AI answers by drawing on content created by online communities, who should really hold the rights to the value generated: the platforms hosting the conversations, the people creating them, or the companies turning them into language models?
Platforms push for control and compensation, creators want recognition, and AI companies want free access.
So it goes back to the old discussion: maintaining a forum is usually a labor of love. Reddit provides the infrastructure, but moderation is done for free by passionate volunteers.
Reddit should have a mutually beneficial relationship with its users and moderators. It could. But LLMs are parasitic. They expect free access and the monetization that comes with it without giving anything back. Lawsuits are currently underway on exactly this question: who holds the rights to the value generated?
Reddit talks about “data laundering” and accuses Perplexity of bypassing blocks to extract content without authorization. But isn’t this proof that generative AI rests on a huge gray area between copyright and freedom of information?
Yes. 100%. No reservations. That is the answer. This is not laundering, it is theft.
What happens when you stop feeding the beast? It seems these online platforms: search engines, social media and others keep eroding the margins of the content middle class.
This is not sustainable and will lead to two extremes: a dead internet inhabited by bots and small closed pockets of human activity that will be impenetrable to search crawlers.
Myriam, when artificial intelligence chooses which ads to show and how to evaluate their performance, how much room is really left for human marketing and independent optimization skills?
Interesting question. AI frees marketers from routine tactical work, but the real differentiator is strategic storytelling and data interpretation. Tools can automate and optimize, but only human beings can create narratives that truly resonate and spark curiosity.
Advertisers spend 80% of their time on optimization hoping to reach that last 20% of extra budget optimization. The cycle of online activity for advertisers usually works like this:
- you find a tactic that works;
- you optimize it to the bone;
- everyone copies it;
- it becomes awkward for humans/algorithms to adapt;
- and you wake up to do it all over again, again and again and again.
Sometimes, in short, building something new is better than perfecting something old.
I think over-optimization tends to make everything look the same.
The Von Restorff effect tells us that people only notice what stands out. You cannot stand out by doing exactly what everyone else is doing, only 2% better. If you want to optimize something to the extreme, rely on machines. If you want to create something new that works and makes you stand out, rely on human beings.
How can users trust AI-powered browsers like ChatGPT Atlas when cybersecurity experts warn they could be vulnerable to prompt injection attacks capable of stealing personal data, draining bank accounts or even manipulating users’ actions directly in the browser?
Well, they can’t! It is a cybersecurity nightmare.
AI-powered browsers are far behind classic browsers in terms of security, and users are exposed to risks. I wish I had something better to say about it, but I don’t.
It is a way for these companies to make money. They want clickstream data and they need a browser and an extension to get it instead of paying for it. In this case, we really are the product.
What we get in exchange for our data is a poor deal. They need to build a better browser!
Right now they are building on borrowed infrastructure. By relying on the Chromium framework, browsers like Comet and Atlas are essentially developing new features on an infrastructure they do not fully control.
This means their core innovations depend on the policies, updates and technical constraints set by the maintainers of Chromium and by Google, making their position inherently dependent on someone else’s foundations. These companies need to offer better products if they want to see mass adoption.
Okay, but if artificial intelligence filters out everything that looks ambiguous or suspicious, isn’t there a risk it will also penalize content that is authentic but complex or unconventional?
It filters more than suspicious things. We are already dealing with this.
It flattens nuance, it averages everything out: faces, music, interior design, content. Homogeneity is pervasive and it dulls the mind. There is a term for it: blanding.
Brands follow a uniform formula in design, typography and color palettes. Slogans get the same treatment. We are converging into one big pile of bland, cliché-ridden aesthetics.
Yes, TikTok is showcasing a myriad of niches and aesthetics, but in real life we are stuck in the era of mediocrity. Let me give you an example: Airbnb International is now a design trend for homes. They are building generic condos. Cars all look the same.
It is a proliferation of copy/paste design. The “Instagram face” trend, driven by injectable “touch-ups”, digital enhancements (filters) and cosmetic techniques, would be alarming if it weren’t so boring and sad.
Think about the standardization of movie posters and cinematic clichés across genres, driven by market testing and preferences. Everything is a universe (or multiverse, ed.) now. The dominance of sequels, prequels and franchise films at the top of the box office means original content is becoming rarer.
We are over-optimizing things to please an algorithm and self-censoring before we even create the content we would like to create.
On the other hand, though, it is a cycle: human beings like trends, and they embrace quirks and outliers (those signals that don’t behave like the rest of the data and show a pattern outside the usual trend).
My divergent thinking is what keeps me in business.
Algorithms pick up signals of what is becoming popular, and not all of these trends can be predicted or created. This is where the concept of differentiation becomes essential.
When AI-powered answer engines get it wrong four times out of five and cite non-existent sources, who will really pay the price of these errors? Will companies have to start defending their reputation from the mistakes of the AI that talks about them and on their behalf?
Yes, when AI-powered engines hallucinate and cite non-existent sources, features and prices, it is mainly brands that pay the price in terms of reputation.
You risk undermining trust in the brand without even realizing it, especially when incorrect information is amplified by popular LLMs like ChatGPT.
AI brand drift, where outdated, incorrect or out-of-context information forms a “shadow brand” outside your control, silently distorts user perception over time.
That is why in my talks and articles I stress the importance of proactive AI visibility audits, continuous monitoring and quick responses. No marketer feels comfortable letting AI tell a brand’s story unsupervised.
Is the Internet really dead? Not as long as human beings populate it with intelligence
As you have read, the conversation with Myriam Jessier touched on some essential points — the role of artificial intelligence, the fragility of LLM-based browsers, the gray areas of copyright, the distortions of algorithms — but what strikes most is the depth with which two often oversimplified phenomena emerge: blanding and the so-called “dead internet” theory.
Blanding is a signal: when everything tends to look the same, the space for those who know how to stand out becomes more valuable.
Standardization aims for safety and predictability of results, even in fields as precarious as search.
Recognizing it as a drift, however, opens the way to a more conscious approach, one in which differentiation becomes an intentional choice again rather than a side effect.
As for the “dead internet”, the image may seem extreme, but it contains a useful element: it reminds us that the web stays alive only if communities, original ideas and authentic voices continue to exist and to produce content that algorithms cannot imitate.
It is not a destiny already written: it is a shared responsibility.
As long as someone creates, discusses, experiments and puts themselves out there, the internet stays alive.
And our work, moving between search engines, AI and editorial strategies, can help keep it that way.
This interview proves once again how essential it is to keep a critical, curious and above all human perspective in an ecosystem dominated by technological speed.
A sincere thank you to Myriam for sharing precious insights and perspectives that enrich the debate and push us to think beyond appearances.
See you in the next episode of SEO Confidential.
