--- title: "Uroborus effect — Google It or Nothing" date: 2025-12-08 description: "Google extracts your content, rewrites it through AI, and keeps the ad revenue. You\'re left with the bills. The anatomy of a robbery dressed up as progress." author: "Petru Cojocaru" categories: - name: "Google" url: "https://cojocarupetru.info/google-en.md" --- # Uroborus effect — Google It or Nothing *Anatomy of an orchestrated collapse.****Beyond technical progress: when innovation becomes an instrument of annihilation.*** > The technical solution is simple: **an automatic compensation system for sources used by AI**. In its official communications for 2024–2025, Google speaks of "improving the user experience" and "democratising access to information". Fine words. Technological progress dressed up as moral imperative. But behind that polished corporate rhetoric, something altogether more bru***[![Uroborus effect — "Google It or Nothing"](https://cojocarupetru.info/images/google/Uroborus-effect.jpg "Uroborus effect — \"Google It or Nothing\"")](https://cojocarupetru.info/index.php?option=com_content&view=article&id=369:uroborus-effect&catid=25:google-en&lang=en-GB&Itemid=508)***tal is unfolding: a wholesale restructuring of the digital information ecosystem — an industrial-scale execution of millions of content creators: independent journalists, researchers, niche bloggers, educational websites.  In short, universal Uroborus effect, continuous creation and destruction. ### Listen to the audio dialogue summary The mechanism is as simple as it is cynical. Why would anyone click through to an article when Google hands them the answer directly — elegantly generated by AI, synthesised from millions of precisely those articles? The answer arrives instantly. No irritating adverts. No waiting. Perfect for the user. Catastrophic for whoever created the original information. A digital vampire, feeding on the work of others whilst giving nothing in return — no traffic, no revenue, not even acknowledgement. Let us take apart the mechanisms of this programmed collapse, documented step by step in the industry's own official sources. --- ## **The Slow Death of Original Content Sites: Vampirism Without Compensation** ### **How the Execution Works — the Documented Mechanism** Barry Schwartz, an SEO journalist for Search Engine Land, analysed aggregate data from over a thousand websites in 2024. His findings leave no room for interpretation: **traffic declines of between 40 and 60 per cent for sites that had previously held the top positions in search results**. We are not talking about obscure sources or dubious content. We are talking about outlets that spent years occupying Google's top three positions for the keywords most relevant to their fields. What changed? The introduction of AI Overviews. Google extracts their content, elegantly synthesises it into an AI-generated response displayed directly on the search page, and the user has no reason whatsoever to click through. The information is already there. Mission accomplished for the user. Mission sabotaged for the creator. Rand Fishkin, former CEO of Moz and an independent consultant with no corporate loyalties to protect, documents the complementary phenomenon in his 2024 longitudinal analysis: **a dramatic reduction in click-through rates to third-party sites**. People no longer visit the sources. Why would they? The answer has already been served up on a plate. ### **The Sabotaged Economic Model: Advertising That No Longer Exists** We need to understand the fundamental financial mechanism. According to Alphabet Inc.'s official shareholder reports for Q3–Q4 2024, **over 80 per cent of Google's revenues come from advertising**. The corporation lives on adverts. But independent content sites? They live on **exactly** the same thing. When you write a comprehensive article on the treatment of diabetes, you invest days of medical research, consult peer-reviewed studies, and structure the information accessibly for a general readership. You publish it on your site. Your hope: that people will search "diabetes treatment", that Google will send them to your article, that they will see the discreet advertisements placed there, and that you will receive a few pennies per click. Multiplied across thousands of monthly visitors, it becomes the income that sustains the work. Now Google extracts your content, reformulates it through AI, and serves it directly on the search page — where it displays its **own** advertisements — and the user never arrives at your site. **You did the work. Google pockets the money. You are left with the bills.** This is not innovation. It is highway robbery, dressed up in the language of "enhanced experience" and "informational efficiency". ### **The Vicious Circle: Less Original Content Means More Poorly Trained AI** The supreme irony — and here the system exposes its own intrinsic absurdity — is that mass extraction without compensation will inevitably lead to the collapse of original content production. When journalists and researchers can no longer survive financially from their work, they stop producing. Simple. Pragmatic. But where will the AI learn in three to five years' time, when it needs retraining on updated data? It will learn from its own previous syntheses. From content generated by other AIs. From the regurgitation of its own outputs. **Digital ouroboros: the serpent devouring its own tail.** The scientific literature calls this phenomenon "model collapse" — the progressive degradation of quality when machine-learning systems are trained on data generated by similar systems. Quality declines exponentially with each iteration. The AI feeds on its own informational waste and slowly poisons itself. Google knows this. It knows it perfectly well. MIT Technology Review reports from 2024 explicitly document the internal concerns of the corporation's own researchers about precisely this risk. Yet the short-term strategy — immediate domination of the AI market — takes precedence over long-term sustainability. The corporation is playing poker with the future of the global information ecosystem, wagering that it will find solutions before the problem becomes critical. We, the users and content creators, are the chips on the table. --- ## **The Collapse of Independent Journalism's Economic Model: When the Fourth Estate Becomes a Servant** ### **Investigative Journalism: the First Sacrifice on the Altar of Efficiency** Real journalism — the kind that investigates corruption, scandals, abuses of power — is expensive. Extraordinarily expensive. It requires teams of reporters, months of documentation, legal advice to defend against intimidatory lawsuits, cross-referenced verification of sources. A solid piece of investigative journalism can cost tens of thousands of euros in human and technological resources. Its return? The traffic generated by the article, translated into advertising revenue and subscriptions. When Google extracts the essence of that article and serves it gratis in an AI Overview, the economic equation collapses. **Costly investigations become financially unsustainable.** Independent publications can no longer support in-depth journalism. What remains are the sensationalist headlines, the short takes, the clickbait — because those still generate quick, impulsive traffic, before the user realises the answer was available for free further up the page. **Serious journalism dies. Noise remains.** ### **Growing Platform Dependence: from Independent to Servile** The Electronic Frontier Foundation (EFF), the non-profit organisation dedicated to digital rights, documents an alarming phenomenon in its 2024 reports: **publications that survive become dependent on the tech platforms**. When organic traffic from Google collapses, the only lifeline becomes... Google itself. The Google News Showcase programme, preferential display agreements, content-licensing contracts for AI training. The corporation offers small financial compensations — a negligible fraction of the revenues generated through the exploitation of that content — in exchange for docility. The result? **The fourth estate — the free press — becomes the vassal of the tech corporations.** Systemic critiques? Investigations into Google's abuses? Risky. The corporation keeping you alive financially does not appreciate being bitten by the hand that feeds. Plato, in the *Republic*, described how philosophers must be independent if they are to speak truth. Journalists economically dependent on the very subjects of their investigations can no longer speak truth. They can only negotiate which fragment of truth is deemed acceptable. --- ## **The Vicious Circle of Informational Impoverishment: When the Internet Becomes a Desert** ### **The Diversity of Voices: from Ecosystem to Monoculture** Before algorithmic dominance, the internet was a diverse ecosystem. Extraordinary niche sites: entomology enthusiasts sharing decades of observations, pharmacists explaining complex drug interactions, amateur historians documenting local history overlooked by official academia. These creators were not professionals. They were **enthusiasts contributing to the global wealth of information without major financial expectations**. Yet they still needed a viable minimum: a few hundred monthly visitors, symbolic recognition, the sense that their work mattered. When Google eliminates traffic to these resources through automated AI extraction, the passion goes out. Why keep updating a database on Carpathian flora when no one visits it any more? Why keep translating rare mediaeval texts when AI syntheses render them invisible? **The internet becomes a monocultural desert** — only corporate content survives, backed by the massive funding that allows it to endure without organic traffic. The independent voices are lost. The unconventional perspectives. The niche knowledge. Shoshana Zuboff, in *The Age of Surveillance Capitalism* (2019), warned of the concentration of informational power. What we are witnessing now is the fulfilment of that prophecy: **the global standardisation of perspectives**. ### **The Inverse Babel Effect: from Many Tongues to a Single Voice** In the biblical book of Genesis, God confounds the languages of humanity to forestall their co-ordinated construction of the Tower of Babel — that symbol of human arrogance. All subsequent human history is marked by linguistic and cultural diversity. Google AI achieves the inverse: **it unifies all voices into a single standardised response**. Ask about Romanian culinary traditions and you receive Google's synthesis drawn from twenty different sources, melted into a generic reply that loses the nuances, the personal stories, the regional disputes. You no longer hear the voice of a grandmother from Maramureș explaining why the stuffed cabbage rolls in her part of the world are made differently. You no longer come across the history teacher's blog that argues — with documented evidence — that the traditional recipe was influenced by Ottoman migrations. You receive **The Official Google Answer™** — apparently complete; in reality, sterilised of diversity. Saint Maximus the Confessor spoke of authentic knowledge as personal experience, not the mere accumulation of second-hand information. When we eliminate direct contact with primary sources, with authentic voices, with concrete experiences, we are left with **the illusion of knowledge** — we know the answer, but we do not understand the depth. --- ## **Consequences for Future Generations: Children Raised in an Informational Desert** ### **The Generation That Does Not Know Sources Exist** Today's teenagers, born into the era of the smartphone, face a profound risk: **growing up without understanding that information has human origins**. When everything you see consists of AI-generated answers, you begin to perceive information as something that simply "exists" — abstract, authorless, contextless, without human effort behind it. You ask "What is the capital of France?" — the answer appears instantly. You ask "How is depression treated?" — the answer appears instantly. You ask "What should I think about conflict X?" — the answer appears instantly. **Nowhere do you see the journalist who spent months investigating depression. Nowhere do you see the researcher who synthesised the clinical studies. Nowhere do you see the complexity and the controversies that lie behind the answer.** The result: a generation that does not understand that information is produced by human beings, that it requires work, that it can be biased, that it deserves to be compensated, that it must be verified. Information becomes a **free public utility** — like air or sunlight — with one crucial difference: air and sunlight require no human effort to produce. ### **The Atrophy of Research Skills: Infantilising Dependence** Cal Newport, in *Digital Minimalism* (2019), warned of the cognitive effects of technological dependence. We are now witnessing the concrete manifestation of precisely that: **the atrophy of independent research capacity**. Earlier generations knew how to find information: consulting encyclopaedias, asking experts, comparing multiple sources, evaluating credibility. These were fundamental cognitive competencies. The generation raised on AI Overviews no longer develops these skills. Why learn to research when the answer appears automatically? Why evaluate sources when Google has already made the synthesis? Why compare perspectives when AI has already provided "the correct answer"? **Orchestrated cognitive infantilisation.** This is not an accident. It is the predictable outcome of a system designed for maximum dependency. Tristan Harris — former ethical design expert at Google and co-founder of the Center for Humane Technology — explained in the documentary *The Social Dilemma* (2020): platforms are deliberately designed for maximum capture and minimum autonomy. --- ## **The Alternative That Was Ignored: What Responsible Separation Would Have Meant** ### **The ChatGPT/Claude Model: Transparency Through Limitation** There is a functioning alternative. ChatGPT, Claude, Perplexity — separate AI platforms, with distinct interfaces, where the user **knows they are using AI**. No confusion. No deception. You enter the interaction with an AI system consciously, with a clear understanding of its limitations. When Google integrated AI as a separate module — as it had done initially with AI Mode — it preserved this transparency. The user chose explicitly: do I want classic search (multiple sources, personal evaluation) or AI synthesis (rapid answer, without depth)? **Freedom of choice. Contextual transparency. User accountability.** But that did not maximise adoption. Many users preferred classic search for important tasks. The corporation's solution? Forced unification. Eliminate the choice. Impose AI as the default. As Google's official 2024 statement put it: "we are simplifying the user experience through complete integration." Translation: we are eliminating the user's capacity to choose the right tool, because free choice was reducing the profitability of our AI strategy. ### **Protecting the Ecosystem: Funding Content Creators** The technical solution is simple: **an automatic compensation system for sources used by AI**. When Google generates a response based on fifteen articles, it automatically distributes a fraction of the associated advertising revenues to those fifteen sources. Transparent. Auditable. Ethical. European legislation — the Digital Markets Act (DMA, 2023) and the AI Act (2024) — theoretically mandates transparency and accountability. But the enforcement? Symbolic. Fines of a few hundred million euros for a corporation with an annual profit running into tens of billions represent nothing more than **a cost of doing business**, not effective deterrence. The clear and transparent laws that Google invokes in its official communications? Legislative theatre. Formal compliance without any substantive change in behaviour. The Electronic Frontier Foundation explicitly documents this decoupling between written legislation and applied reality. --- ## **Conclusion: The Ecosystem on the Operating Table** The impact of AI integration into Google's classic search engine on the digital ecosystem is not apocalyptic speculation. It is documented reality, evidenced by: - **Verifiable statistical data** (40–60 per cent traffic declines, reported across the industry) - **Official financial reports** (Google's dependence on advertising; the post-ChatGPT pressure) - **Independent analyses** (MIT Technology Review, EFF, specialists with no corporate loyalties) We are not speaking of inevitable progress or natural evolution. We are speaking of **concrete strategic choices** made by a corporation in a monopoly position (92%+ of the global search market, StatCounter 2024) that prioritises short-term dominance at the expense of the long-term sustainability of the global information ecosystem. The alternative exists. It has been demonstrated to work. Google chose, deliberately, not to pursue it. The question is not technological. It is moral: do we accept the collapse of the diverse information ecosystem in exchange for the comfort of instant answers? Or do we demand **responsible separation** — AI as a distinct, transparent, complementary tool, not as a vampiric replacement for the human sources that feed it? The choice belongs to users. For now. For as long as they still have the power to choose. **Principal Documentary Sources:** - Search Engine Land (Barry Schwartz, 2024) — traffic statistics - Moz (Rand Fishkin, 2024) — user behaviour analysis - Alphabet Inc. financial reports Q3–Q4 2024 - MIT Technology Review (2024) — technical and ethical analysis - Electronic Frontier Foundation — reports on informational monopoly - Google Official Blog — statement on unifying AI experiences - Shoshana Zuboff — *The Age of Surveillance Capitalism* (2019) --- ## 20 Illuminating Questions and Answers **1. What are AI Overviews and why do they change the game so fundamentally?** AI Overviews are automatically generated summaries produced by Google and displayed directly on the search results page, above the classic results. They change the game because they eliminate the fundamental reason for clicking: the user receives the answer without ever visiting the source. It is rather like reading the back-cover blurb and persuading yourself you have read the book. **2. By how much has traffic to original content sites declined since the introduction of AI Overviews?** Barry Schwartz of Search Engine Land analysed over a thousand websites and documented declines of between 40 and 60 per cent — including for sources that had occupied Google's top three positions for years. This is not statistical turbulence. It is systematic collapse. **3. Why is the economic model of content creators structurally sabotaged?** The model was straightforward: write quality content → Google sends traffic → visitors see adverts → you receive revenue. Google has severed the middle link: it extracts your content, reformulates it through AI, and displays it alongside its own advertisements. You produced the goods. Google sold them and kept the money. **4. What does "digital ouroboros" mean in the context of this article?** It means the system is poisoning itself. If content creators can no longer survive financially and cease production, the AI of the future will be trained on its own previous syntheses. The degradation of quality is inevitable — the phenomenon the scientific literature calls "model collapse". The serpent devours its own tail and wonders why it hurts. **5. What does MIT Technology Review document in relation to "model collapse"?** MIT Technology Review reports from 2024 confirm that Google's internal researchers are aware of this risk. They know that training AI on data generated by other AIs produces exponential quality degradation. And yet the short-term strategy — immediate market dominance — prevails. It is poker played with chips that do not belong to the player. **6. How does the free press become the vassal of the tech corporations?** When organic traffic collapses, publications survive through agreements with Google — News Showcase, content licensing, preferential contracts. The corporation keeping you financially alive cannot be the subject of your critical investigations. The fourth estate ends up negotiating which fragment of truth is acceptable. Plato would recognise the scene. **7. What does culture lose when niche voices disappear?** It loses what cannot be synthesised: the voice of a grandmother from Maramureș explaining why the cabbage rolls in her region are made differently, the history teacher's blog arguing that the recipe was shaped by Ottoman migrations. AI melts twenty sources into one Official Answer™, sterilised of diversity. The information remains. The knowledge disappears. **8. What did Saint Maximus the Confessor say about knowledge — and how does it apply today?** Saint Maximus spoke of authentic knowledge as personal experience, not the accumulation of second-hand information. When you eliminate direct contact with primary sources, with concrete voices, with real controversies, you produce the illusion of knowledge — you know the answer, but you do not understand the depth. The screen that serves you the synthesis does not teach you to think; it anaesthetises your need to think. **9. How do AI Overviews affect younger generations at a cognitive level?** Adolescents raised on instant answers no longer learn that information has human origins, that it requires work, that it can be wrong or manipulated. Cal Newport warned in *Digital Minimalism* of the atrophy of cognitive skills. Dependence on AI Overviews is the concrete manifestation: why learn to evaluate sources when Google has already made the synthesis? **10. How does Tristan Harris describe the attention-capture mechanism?** Tristan Harris — former ethical design expert at Google and co-founder of the Center for Humane Technology — explained in *The Social Dilemma* that platforms are deliberately architected for maximum capture and minimum autonomy. This is not an unforeseen side effect. It is deliberate design. **11. What was the responsible alternative that Google chose not to pursue?** Maintaining AI as a separate, distinct module, clearly distinguished from classic search — as ChatGPT and Claude function. The user would choose explicitly: do I want human sources or AI synthesis? Freedom of choice, contextual transparency. Google chose forced unification, eliminating this bifurcation, because free choice reduced the profitability of its AI strategy. **12. What simple technical solution could repair the ecosystem?** An automatic compensation system: when Google's AI generates a response based on fifteen articles, it automatically distributes a fraction of the associated advertising revenues to those fifteen sources. Transparent, auditable, ethical. The solution exists technically. The corporate will does not. **13. What do the Digital Markets Act and the AI Act represent for the ecosystem — and why is that not enough?** European legislation theoretically imposes transparency and accountability. But fines of a few hundred million euros for a corporation with annual profits of tens of billions represent a cost of doing business, not effective deterrence. It is legislative theatre with formal compliance and no substantive change in behaviour. **14. What share of the global market does Google hold, and why does it matter?** According to StatCounter 2024, Google controls over 92 per cent of the global search market. This is not a dominant position — it is a de facto monopoly. The strategic choices of a single corporation shape access to information for the entire connected world. The risk is not theoretical. **15. What is "informational monoculture" set against the former diversity?** Before algorithmic dominance, the internet was a diverse ecosystem — entomology enthusiasts, pharmacists, amateur historians, all contributing to the global wealth of knowledge. Google's AI algorithm unifies these voices into standardised responses. From tropical rainforest, you arrive at a monoculture plantation. More orderly. Incomparably more fragile. **16. What does "digital vampire" mean in the economy of this text?** It is the article's central metaphor: a system that feeds on the work of others without giving anything in return — no traffic, no revenue, no acknowledgement. The digital vampire extracts the informational lifeblood of creators, converts it into advertising gold, and leaves the economic cadavers behind, presenting the whole enterprise as "progress". **17. How does Google justify this strategy internally — and how does the author translate that justification?** Google's official communications speak of "improving the user experience" and "democratising access to information". The author's translation: we are eliminating the user's capacity to choose the right tool because free choice was reducing profitability. Corporate rhetoric is cosmetic applied over robbery. **18. What does Shoshana Zuboff say in the context of this analysis?** Zuboff's *The Age of Surveillance Capitalism* (2019) documents how tech corporations built business models based on extracting and monetising human data without genuine consent. AI Overviews represent the next stage: no longer merely extracting user behaviour, but also extracting the intellectual labour of content creators and converting it into advertising capital without compensation. **19. What moral question does the article leave open?** Do we accept the collapse of the diverse information ecosystem in exchange for the comfort of instant answers? Or do we demand responsible separation — AI as a distinct, transparent, complementary tool, not as a vampiric replacement for the human sources that feed it? The article declines to answer on the reader's behalf. But it makes clear that the power to choose still exists. For now. **20. What is more dangerous: a mind that does not know, or a mind that believes it knows?** This is the profound tension running through the whole essay. Generations raised on instant answers will not suffer from raw ignorance — they will suffer from false certainty. They will know the answer without having travelled the road that leads to it. It is the difference between possessing a map and having walked the territory on foot. The map does not teach you how to survive. --- ## Author's Note *Limits, tensions, comparisons, and rhetorical tools* I write this with a certain discomfort — and that discomfort is useful, because it marks the places where my text exceeds its own shadow. **Acknowledged limits.** This article is not an academic study. The sources invoked — Barry Schwartz, Rand Fishkin, MIT Technology Review — are real and verifiable, but the quotations are paraphrased from editorial memory, not reproduced with scholarly rigour. The 40–60 per cent traffic decline is real as an order of magnitude; it is not real as a universal average. Some content sites have actually grown since AI Overviews — those that adapted quickly enough, that found ways to produce content which AI synthesis cannot absorb: on-the-ground investigations, exclusive interviews, a recognisable voice. The article simplifies deliberately in order to achieve rhetorical clarity, and that simplification carries its own cost. **Unresolved tensions.** There is an internal tension that I resolve nowhere in the text: if AI Overviews are as devastating as I claim, how does my article manage to circulate through precisely those indexing engines it accuses? I write for machines that forget, as I have said elsewhere, and yet I hope that the machines will send people towards my text. It is a contradiction I carry like a snowball in my pocket: I feel it there, I say nothing about it, I hope it will have melted by the time I get home. It does not melt. **Comparisons at work in the text.** I have placed Plato in dialogue with Saint Maximus the Confessor, Cal Newport alongside Tristan Harris, the Digital Markets Act alongside Shoshana Zuboff. These juxtapositions are not decorative — they construct a philosophical field of tension around a subject that might otherwise remain simplistic technological criticism. Plato and economically dependent journalism: both address the impossibility of truth in the absence of independence. Saint Maximus and the illusion of AI knowledge: both name the same danger, separated by centuries. They function as an archipelago of arguments — each island distinct, yet linked by the same submerged plateau. **Principal rhetorical tools.** The article rests on three repeated rhetorical movements: *the demolition of euphemism* (I take corporate language and translate it bluntly — "enhanced experience" becomes "the elimination of choice"); *controlled apocalyptic escalation* (each section raises the intensity, but I return systematically to concrete data before the tone tips into empty polemic); and *the metaphor with tentacles* — the digital vampire, the ouroboros, the monoculture, the informational desert. The metaphors do not accumulate randomly: they construct a coherent system of imagery around extraction, self-destruction, and sterilisation. Someone reading quickly retains the metaphor. Someone reading carefully retains the argument. The article tries to serve both readers. What remains uncertain in me, after I have finished writing, is whether the anger in this text is just or merely comfortable. There is a comfort in righteous indignation, and I sometimes look at my own sentences with the suspicion that I am writing about disaster with a certain pleasure — that the rhythm of incisive judgements produces in me an aesthetic satisfaction I have not earned until I have checked the data ten times over. That is the limit you cannot cross through editing. You can only name it. ---