AI's Tectonic Shift : The Transition from a Link-Based Ecosystem to an AI Information Consumption Model

1. The Transition from Link-Based to AI Information Consumption

I feel it in my bones that Google search, which emerged like a comet in 1998, dominating the market for 27 years and capturing over 90% of the global search market, is taking a backseat. At the same time, I also feel that the seismic shift being brought about by AI is significant. Right now! The ecosystem that thrived based on links is transitioning to an AI information consumption model.

Sometime ago, even I started using ChatGPT, Google AI Studio, Claude, and Perplexity instead of searching on Google first. That's because these AI services find, organize, and summarize the information I want for me. Furthermore, they make it incredibly easy for me to give commands, find what I'm missing, and even what else I need, and organize it all. Now, I only search on Google when necessary, like to verify the facts of information obtained through AI services, or to find the absolute latest information.

2. The Change in the Advertising Business

What business model should the countless websites that lived off advertising—all those search terms, search ads, banner ads—now adopt? The so-called Big Tech companies (AI companies) are gobbling up all the information in the internet world (learning from it) and providing AI services. There's no longer a reason to follow links to destination sites. What kind of business model will Big Tech companies propose for their lucrative advertising business? What kind of symbiotic model will they offer to the countless internet companies, content providers, and individual bloggers, excluding Big Tech?

The traditional search engine presented a list of web pages (links) containing relevant information when a user entered a query, and users obtained the desired information by clicking on these links. Companies including Google generated revenue by displaying ads based on the traffic (pageviews, visitors) generated during this search process. Thus, the search advertising market was the core of the digital advertising market and Google's main revenue source for 27 years. However, AI engines based on Large Language Models (LLMs), such as ChatGPT, Claude, Gemini, Perplexity, and Grok, synthesize and summarize information from multiple sources in response to a user's question (prompt), providing a single, integrated answer. Naturally, these AI answers reduce the need for users to individually click links on various websites to search for information. The tendency for the 'destination' of AI information search to increasingly become the AI engine itself is growing stronger. This is a huge challenge that demands a redesign of the search-based business and advertising ecosystem built over decades.

3. Advertising in the Age of AI

Let's look at the current usage patterns of AI services. While I agree that providing sources with the answer results *does* increase the objectivity of the provided material, it is often merely a supplementary means to confirm the existence of evidence. Again, instances of following links to enter and view the destination website are remarkably low.

Ultimately, to encourage user movement, it will likely be provided in the form of native advertising included within the AI answer, but even that is fundamentally different from search advertising. If information gathering is satisfied within the current chat window through questions and answers, and links are only used for a few specific actions (purchase, booking, etc.), it is highly likely to not generate the traffic of the past. Furthermore, the recommendation algorithms of just a few Big Tech companies are so artificial that it's impossible to tell if it's truly based on genuine needs or if it's artificially induced by a combination of personal information and queries. Whether clicks or impressions generated in that way are truly helpful requires time to see. It needs to be proven in the market what their value is and how they differ from simple clicks, impressions, and conversions of the past.

4. The Essence of the Threat and Potential Solutions in the AI Age

source : https://blog.google/products/search/ai-mode-updates-may-2025/

More than anything, the most significant threat in the AI transition era is the decrease in traffic. AI crawling engines will buzz around actively, gathering and learning information from the web. Indeed, non-human traffic, primarily bots, already accounts for a significant portion (estimated at over 80%) of total web traffic. Recent data specifically highlights that OpenAI's GPTBot constitutes about 13% of total web traffic, and combined with other AI crawlers, their volume reaches 28% of total Googlebot traffic. Human-generated traffic to individual sites will decrease even more significantly. This is because users no longer need to visit individual websites directly as they obtain answers through AI services. This problem reduces the number of visitors to websites and decreases pageviews. For this reason, a few Big Tech companies could gain more control and potentially adjust individual revenues with unknown algorithms.

Even though websites provide information, the banner ad revenue earned on those sites is likely to decrease significantly. In particular, web services that provided news or community services may face serious financial problems due to declining advertising revenue. If Big Tech companies fail to create a symbiotic ecosystem, it will naturally weaken content creation motivation. If only a few Big Tech companies try to monopolize everything, it is self-evident that the entire internet ecosystem will weaken.

Ultimately, content creators will someday come to grapple with the serious question: 'To what extent should we allow AI crawling?' How much should they disclose to be recognized for their expertise by AI services and still generate *some* traffic? Or should they block AI learning altogether and try to make users type the website address directly or search for it and find them? This is by no means an easy problem.

Let's look at this from another angle. The existing open and distributed web ecosystem may increasingly head towards closure. This is because one easy way to solve this problem is the shift from the advertising business to content monetization. From using content for free in exchange for seeing banners or ads in the past, the tendency to transition to a closed, paid model based on expertise may strengthen.

This change shows it's a key variable that will determine the future of the internet information production-consumption ecosystem. If an open and mutually beneficial cooperative model is not established, the internet information production-consumption ecosystem is likely to shrink. Closing wide-open doors, specializing, and selling dearly is not a panacea.