We are living through an ecological catastrophe. Only this one isn't happening in the Amazon rainforest, but in the digital ecosystem of the internet.
❯ Prologue: A New Predator in the Savannah
Imagine a pack of apex predators suddenly unleashed upon the African savannah. Not ordinary lions—but creatures that run faster than cheetahs, see better than eagles, never sleep, and know no mercy.
What would happen to the ecosystem?
The herbivores, the slow grazers, would be the first to disappear. Only those who could run exceptionally fast or hide in burrows would survive. The mid-level predators would die out, with nothing left to hunt. The entire food chain would restructure itself around the new dominant species.
The very same thing is happening to the internet on a global scale, right now. AI assistants are those apex predators, bursting onto the digital savannah. And they are radically and successfully reshaping the entire ecosystem in their own image. Only instead of antelopes and zebras, it's information sites that are going extinct. Instead of hyenas and jackals, it's content aggregators that are disappearing. And in place of a once-rich ecosystem of knowledge, a digital desert of entertainment is all that remains.
❯ The Source of the Catastrophe: The Battle for the Last Resource
In any ecosystem, there is a limited resource that drives the central conflict. In the savannah, it's food and water. On the internet, it's the user's time—human attention.
Before AI, the ecosystem worked like this:
A user searched for information → A search engine offered options → Websites competed for clicks → The user chose a source → They read, compared, and drew conclusions.
This was a complex, multi-layered food chain. Each link received its share of attention. Now, the chain has been shortened to just two links:
A user asks a question → AI provides a ready-made answer.
That's it. The other links in the chain became obsolete.
❯ The First Victims: Who is Already Feeling the Pressure
Wikipedia: A Giant with Feet of Clay
I can't remember the last time I visited Wikipedia. Citing it may be out of fashion now, but Wikipedia remains the crown jewel of the internet ecosystem—humanity's collective mind. But why go there if GPT can summarize any article in seconds? And do so not in a dry manner, but adapted to your level of understanding? Sure, it might miss nuances and invent authors, citations, and concepts, but who's going to check?
The numbers speak for themselves: according to a recent study, Wikipedia's traffic has fallen by 23% in three years, an incredibly sharp decline for a top site. Researchers directly link this to competition from AI tools. Note that in March 2025, chatGPT.com attracted 500 million more visits than Wikipedia. Who will spend hours writing an article that no one will read? AI is sucking the knowledge from this collective brain but giving nothing in return. Wikipedia is turning into a zombie—formally alive, but with the life slowly draining out of it. The very source that feeds AI is itself slowly dying of exhaustion.
Professional Forums: The Stagnation of Expert Communities
Stack Overflow, specialized forums for programmers, medical communities—all these places where experts shared knowledge and debated subtleties are losing their purpose. The number of questions and answers on Stack Overflow in April 2025 fell by 64% compared to April 2024. Why search for an answer on a forum, reading through half a page of discussion, when AI can provide a concentrated solution instantly?
"Vibe-coding" is an alluring mirage for amateurs. Newcomers build a startup on the fly without understanding what's happening inside the code. AI stacks black box upon black box, leaving others to figure it out if something breaks.
Experts have no one to pass their knowledge on to. Newcomers don't ask questions—they ask AI. Decades of accumulated collective wisdom are becoming a museum exhibit. This is especially ironic because it was on these very forums that AI once learned. And now, it is strangling its own parents.
News Sites: Under Threat of Automation
News sites are already feeling the pressure. People still want to know what's happening in the world right now. But accessing news agency websites is no longer essential.
AI can already compile news summaries from dozens of sources. What will happen when it learns to do this in real time? Why read five different articles about an event when AI can give you the full picture in a minute?
News sites will find themselves in the position of farmers growing crops for someone else who then sells it under their own brand. Analytical articles won't save them; AI can already write those. Nor will interviews with politicians—AI will find them on Twitter. The only exception might be top-level exclusives, but that's a very niche market.
It's no wonder that major media conglomerates are filing multi-billion dollar lawsuits against AI creators, accusing them of unlawful content use. They were the first to realize they were being turned from partners into fodder.
❯ The Survivors: Who is Adapting to the New Ecosystem
TikTok and YouTube: The Kings of Entertainment
Interestingly, entertainment platforms aren't just surviving; they're thriving. TikTok, YouTube, and Instagram have no fear of AI. On the contrary, they use it.
Why? Because they sell emotions, not information. Their product isn't answers to questions, but a way to kill time, to laugh, to be amazed.
For now, AI can't create the perfect funny cat video or a captivating travel vlog; it lacks charisma and empathy. But even here, the first niche of AI influencers is already emerging...
Reddit and Community Forums: Oases of Human Connection
Social platforms are also surviving, but for a different reason. People don't go there for information, but for interaction. To argue, to share memes, to feel a sense of belonging to a community.
AI can answer the question "how do I fix a leaky faucet?", but it can't yet replace the pleasure of a witty comment or participation in a pointless debate.
However, there's a catch here too. What happens when half the commenters are bots? When AI learns to create memes and jokes indistinguishable from human ones? Will people eventually grow tired of machine-generated participation and artificial conflicts?
Specialized Services: Niches for Survival
Sites that provide unique services are surviving: banks, e-commerce stores, booking systems, government portals. Anything that requires real-world interaction, not just information. For now, classified ad boards, digital libraries, and dating sites will live on.
But even here, AI is nipping at their heels. Chatbots that can book a table or arrange insurance are already appearing. It's one small step from aggregating classifieds to accepting them directly within an AI app, which then pushes your ad to the perfect buyer. Or even subtly convinces them they need to buy it.
AI translators and AI writers are already cluttering the literary field. It will be a sad day when people get used to this—for now, AI produces only a surrogate, a chewing gum for the brain that simulates the thought process.
Premium Knowledge: Elite Sanctuaries
A new type of survivor is emerging: closed ecosystems of knowledge. Paid communities where real experts share their experience for a fee. Courses, masterclasses, private Slack channels. These are digital sanctuaries, inaccessible to AI crawlers. The knowledge within is protected not just by a password, but by legal non-disclosure agreements.
Imagine: instead of open programming forums, we'll have private clubs for $100 a month. Instead of free YouTube tutorials, premium masterclasses for $500. Knowledge is becoming a privilege once again. But can they fully protect themselves from AI? I doubt it. And it's difficult to assess the real value of information in such closed communities.
The irony is that the internet, which was supposed to democratize access to information, may be returning to a feudal model. Digital monasteries for the chosen few.
❯ The Changing of the Guard: Who Sets the Rules of the Game
A Monopoly on Truth
In the old ecosystem, a person could compare a few sources from a search engine's results and form their own opinion. In the new one, AI itself decides which sources are "authoritative" and which are not. Who decides that The New York Times is more important than an independent journalist's blog? Who determines that a Harvard professor's opinion carries more weight than the observations of a practitioner from a small town? Who decrees that an English-language source is more correct than a Russian-language one?
AI creates an invisible hierarchy of authority. And only those deemed worthy by the algorithm can enter it. This is a new form of censorship—not banning words, but ranking them by importance. The goal will no longer be simply to be useful, clear, and innovative; success will depend on being cited by an authoritative source. And the user, glued to their AI assistant's screen, will never realize how much more diverse the world truly is. While search engines at least leave the potential to see alternatives, with AI, it's far more complicated.
There is another problem: the loss of context. AI rips quotes from articles, ignoring nuances. Medical advice loses crucial disclaimers about contraindications. A legal rule loses its exceptions. A scientific discovery loses the boundaries of its applicability. Imagine a doctor treating patients based on excerpts from textbooks without reading the full chapters. This is precisely what AI is doing to human knowledge—turning it into dangerous fragments.
The Advertising Ecosystem Under Threat
The entire internet economy is built on advertising. Websites earn money by displaying banners. Google gets rich from ads in its search results.
But if people stop visiting websites, who will see the ads? If AI provides ready-made answers, who needs ad listings in search results? Advertisers understand this. The money will flow to where the attention is—to the creators of AI. Google is already integrating ads into its AI Overviews and Bard's responses. ChatGPT is testing sponsored insertions.
The new ecosystem will belong to a handful of AI giants. They will become the sole gatekeepers to the digital world. And when AI evolves into a browser and absorbs all office utilities, the digital infrastructure will collapse, leaving AI as the single point of entry into the digital world for humanity.
❯ The Self-Destruction Paradox: When the Predator Eats Its Own Prey
And now for the most interesting part. AI weakens the very sources it learned from. It's like a predator that has eaten most of the herbivores and is now facing the threat of starvation.
Imagine the internet in 10 years:
Wikipedia is abandoned.
Forums are dead.
News is written by AI.
Bloggers are replaced by AI avatars.
Comments are generated by bots.
What will the next generation of AI learn from? From texts created by the previous generation of AI?
We will witness digital inbreeding—degeneration from breeding with oneself. AI will become increasingly average, formulaic, and predictable. The very "water" of the ecosystem—the strange thoughts, unexpected connections, and human errors that make texts alive—will vanish.
Another threat is also materializing: content entropy. As AI churns out monotonous texts, search engines lose their ability to distinguish quality. Google is already desperately fighting AI spam, but only the most primitive attempts are being filtered out.
The internet is being filled with a gray mass of beautifully written but empty texts. Like a stagnant swamp where all life has suffocated, and which now serves only as a source of peat.
❯ The New Ecosystem: What Replaces the Library
Where are we headed?
The internet is transforming from a "global library" into a "global amusement park." To be honest, only a part of the internet could ever be called a library. Digital junkyards, social spiderwebs with fat SEO spiders, bought-out newspapers, heavily moderated forums clogged with ads, industry portals selling top rankings, porn of all kinds, creative sites full of lunatics... All of this was also the internet.
But even that internet was, and still is, diverse. Everyone could find their own corner. It might not always be approved of by society or your spouse, but you could find a place that was interesting to you. Now, that place will be flooded with monotonous content. Your conversation partners will seem suspiciously like bots. You will visit it less and less, until one day you realize that the endless horizon has shrunk into the small, comfortable window of ChatGPT.
❯ Epilogue: What Awaits Us?
Is this inevitable? Can anything be changed?
Perhaps we need to recognize the value of what we are losing—the value of reading sources, not summaries; of comparing viewpoints; of analysis; of forming our own opinions. The value of subjectivity and necessary human error.
Should we somehow support the "sanctuaries" of the old ecosystem—the places where people still argue, make mistakes, and create unique content? Or maybe a law will be passed for equal access, quotas for website representation in AI sources, or an "AI solidarity tax" will be introduced. But I can't even imagine how such a thing could be implemented.
Or should we accept the inevitable and prepare for a world where knowledge is monopolized, entertainment is algorithmized, and human curiosity is satisfied by a single source?
Then again, what happens to a species that stops hunting for itself, that has its food brought to it ready-made? It doesn't necessarily go extinct. It simply becomes domesticated.