For more than two decades, the web has been structured around an implicit but effective pact: internet users accessed a variety of content—articles, tutorials, videos, forums—free of charge, in exchange for a little of their attention, monetized through advertising. This model, inherited from the press and adapted to the digital age, was based on a relatively stable value chain: content creators (publishers, bloggers, collaborative platforms, etc.) published SEO-optimized pages, search engines such as Google acted as technical intermediaries, and users, by clicking on the links displayed, ensured audience circulation and revenue generation. This organic traffic, often referred to as “free,” was in fact the lifeblood of the industry: it determined visibility, profitability, and ultimately, the viability of the media.
However, this mechanism is now threatened by a technological paradigm shift: the emergence of conversational agents and automated response engines capable of summarizing, rephrasing, or rewriting existing content without redirecting the user to the source. The web is becoming a source of raw material for artificial intelligence that extracts information without compensation. In this new landscape, value no longer lies in the click, but in the immediate response—often generated without the user ever seeing the original page.
This upheaval calls into question one of the economic pillars of the web. This is what we will examine in our July series.
Historically, search engines—led by Google—were the central infrastructure for distributing traffic on the web. They acted as a conduit between internet users and content producers, directing a stream of “qualified” traffic to sites that matched search intentions. This model was based on a relatively stable dynamic: for every two pages indexed, one click was generated. In other words, publishers could reasonably expect that one in two internet users would end up visiting their site after a search.
However, this equation has been profoundly disrupted by the emergence of generative artificial intelligence in the user journey. In 2025, this ratio has fallen to one in six: it now takes six indexed pages to get a single click, according to data reported by Matthew Prince, CEO of Cloudflare. The majority of queries are now answered without the user having to visit any external sites.
The consequence is clear: original content—which is still being produced in large quantities—no longer attracts enough visitors to justify its creation costs. This weakens the entire ecosystem that sustained the open web: advertising agencies, freemium models, affiliate partnerships, and even digital subscriptions based on visibility.
The arrival of artificial intelligence tools such as ChatGPT, Claude, and Perplexity has profoundly changed the way internet users access information. Whereas traditional search engines (such as Google) displayed a list of links to websites, these tools provide complete answers directly, without redirecting the user to the original page.
These AIs work by absorbing billions of web pages to produce “new” texts that are often well-written and easy to read. The problem is that they make extensive use of content created by others (articles, forums, tutorials, etc.), but without crediting or directing visitors back to the source sites.
Matthew Prince (Cloudflare) provides some telling figures: for every click that OpenAI sends to a site, it consumes around 250 pages. At Anthropic (creator of Claude), this figure is up to 6,000 pages for a single click. This means that the vast majority of content is used, but almost no one will read it on its original site.
This is a major disruption: publishers are still creating content, but visitors are no longer coming. AI is becoming an intermediary that prevents the link between the reader and the author. This destroys the very logic of the open web, where clicks used to finance the production of information.
Today, 80% of users resolve 40% of their queries without clicking on a result (Bain & Co, cited by the WSJ). Integrated responses (AI summaries, featured snippets) minimize the need to visit websites, undermining traditional SEO.
Google has begun rolling out a new search format called AI Mode. Instead of its traditional results (the famous “SERPs” – Search Engine Results Pages), the search engine now offers answers generated automatically by artificial intelligence. These answers summarize several sources without direct links to the original sites.
In layman’s terms, users get an instant summary, written by Google’s AI, at the top of the page. In most cases, they no longer need to click on a link to find what they are looking for. As a result, organic traffic—i.e., natural visits from searches—is plummeting, especially for small websites and blogs.
In this new landscape, only the big established players (institutional sites, major brands, Wikipedia) still sometimes appear as references in the summaries. For everyone else, visibility is plummeting.
In response to this change, a new discipline is trying to emerge: Answer Engine Optimization (AEO) or Generative Engine Optimization (GEO). The idea is to optimize content not to appear in search links, but to be included in the response generated by AI. This requires a very direct, structured style, with clear data and well-defined questions and answers. But even with these efforts, there is no guarantee that a site will be cited or that the user will click on it.
When visitors stop coming, revenue evaporates. However, the vast majority of content creators—journalists, bloggers, comparison sites, reviewers, or review sites—depend on traffic to survive. Fewer clicks mean less revenue, and therefore less incentive to produce content.
The model is based on a simple chain: creator → content → visit → monetization. If the link with the audience is broken, the whole equation collapses. Matthew Prince (Cloudflare) puts it clearly: “If content creators can’t derive value from it, they will simply stop creating. “
The pillars of web monetization are directly threatened:
When AI provides the answer directly, no one clicks anymore, and the publisher becomes invisible in digital exchanges. The risk is systemic: a web without traffic is a web without new content, leading to a rapid depletion of the very material that AI exploits.
The consequences of the rise of generative AI go far beyond the issue of revenue alone. On the one hand, these technologies rely on very costly infrastructure: AI systems must query thousands of pages to produce a response, which causes server, bandwidth, and energy requirements to skyrocket. Sites such as Wikipedia and StackOverflow complain about the massive use of their resources without compensation. As a result, they pay more to feed AI than they receive in aid.
On the other hand, we are seeing a proliferation of content created solely for AI, not for humans. Thousands of automatically generated texts—unreliable, poorly written, often empty—are flooding the web. These are known as “content farms” or “AI-generated garbage.” This promotes misinformation, clickbait headlines, and ultimately undermines users’ trust in the results they find online.
Join us in mid-July for the second part of this article, in which we will explore whether it is possible to reinvent audience flows and financial models. In the meantime, if you have a movie, series, software, or e-book that you want to protect, don’t hesitate to call on our services by contacting one of our account managers. PDN has been a pioneer in cybersecurity and anti-piracy for over ten years, and we’re sure to have a solution to help you. Enjoy reading, and see you soon!
Share this article
