![](\"https:\/\/www.newsbeep.com\/us\/wp-content\/uploads\/2026\/11\/a-262-150x150.jpeg\"\/)
<\/p>\nKey Takeaways:<\/p>\n
The rapid rise of AI agents that search, browse, fill forms and transact on behalf of users is challenging the internet\u2019s foundational assumption that humans are its primary audience, with Cloudflare reporting a more than 1,700 percent increase in AI-agent requests and more than half of its network traffic now non-human.<\/p>\n
This shift threatens publishers by reducing human page views and advertising or subscription revenue, while prompting new controls and payment systems for AI access and raising the prospect of machine-to-machine commerce and information cycles.<\/p>\n
The article concludes that the deeper issue is who controls the intermediary layer between humans and information, as agents increasingly filter, summarize and act, potentially redesigning the web around machine users.<\/p>\n
For three decades, the internet was built around one central assumption: humans would use it. People searched for information, read articles, watched videos, compared products, clicked advertisements and visited websites. Much of the digital economy was therefore designed around human attention. That assumption is now being challenged by the rapid emergence of AI agents.<\/p>\n
AI agents do more than answer questions. They can search, browse websites, compare information, fill forms, manage tasks and increasingly make transactions on behalf of users. In September, OpenAI introduced new autonomous agent capabilities designed to pursue goals across applications, while other companies are developing systems that can plan trips, shop for groceries and manage subscriptions.<\/p>\n
From Search Engines to Machine Users<\/p>\n
Traditional search engines sent people from one website to another. Users opened links, read different sources and decided what to believe or buy. AI systems are changing that process by bringing information together inside the interface itself. The user may receive a synthesized answer without visiting most of the websites from which the information came.<\/p>\n
The next stage is more significant. AI agents can interact directly with websites and perform actions rather than merely retrieve information. Cloudflare now describes these systems as a second audience for the internet and says that AI-agent requests on its network increased by more than 1,700% over the past year. It also reported that more than half of internet traffic on its network is now non-human.<\/p>\n
This does not mean that humans have disappeared from the internet. It means that the web now has another rapidly expanding audience. That audience does not browse in the same way as people do. It does not care about page design, advertisements or scrolling. It is software acting on behalf of people.<\/p>\n
The Crisis for the Open Web<\/p>\n
This development creates a serious economic problem for publishers. If an AI system reads an article and gives its user the essential information, the user may never visit the original website. The publisher loses a page view and potentially advertising, subscription or commercial revenue.<\/p>\n
The pressure is already visible. Chartbeat data reported by Axios showed significant declines in referral traffic from traditional search engines, particularly for smaller publishers. AI-driven discovery has not yet fully compensated for those losses. More recently, Axios reported that traffic to news sites is falling as users move from conventional search and social platforms toward AI interfaces.<\/p>\n
This raises a fundamental question: who pays for the journalism and expertise that AI systems use? If millions of websites provide information to machines but receive little traffic or revenue in return, the economic foundation of the open web could weaken.<\/p>\n
Websites Fight Back<\/p>\n
Website owners are beginning to demand greater control over machine access. Some want AI systems to discover their content because they see new opportunities for distribution. Others want to block automated access because their business depends on human visitors.<\/p>\n
Cloudflare has responded by giving website owners separate controls for search, AI training and AI-agent traffic. It has also introduced a system that can allow publishers to receive payment when their content is used by AI systems.<\/p>\n
This could produce a very different internet. Some websites may remain freely accessible to humans but restrict machines. Others may allow agents to enter only after authentication or payment. The open web could gradually acquire a new economic layer in which information has different prices for different kinds of users.<\/p>\n
Commerce Without the Shopper<\/p>\n
The transformation may become even more visible in online commerce. Today, consumers search for products, compare prices and decide what to purchase. Tomorrow, an agent may perform most of those steps.<\/p>\n
The technology is already moving in that direction. AI-agent companies are developing systems that can plan trips, purchase goods and manage subscriptions. Visa has said transactions initiated by AI agents could become a major financial category, while retailers are experimenting with ways to allow or restrict autonomous purchasing.<\/p>\n
If this model becomes widespread, the traditional online shopper could become less visible. A person might simply state a budget and a requirement. The agent could search hundreds of products, compare them and present a recommendation. The human would make the final decision, but much of the internet activity would happen between machines.<\/p>\n
Who Controls the New Internet?<\/p>\n
The deeper issue is not simply whether AI will use the web. It is who will control the relationship between humans and information.<\/p>\n
Today, people choose what to search for and which sources to examine. In an agent-driven internet, software may increasingly determine which pages to read, which products to compare and which information to present. The intermediary could become as important as the websites themselves.<\/p>\n
This creates questions about transparency, competition and accountability. If an agent recommends one product rather than another, users may not know exactly why. If an agent summarizes several news sources, they may not see the original reporting. The power to filter information could therefore move from search engines and social platforms toward AI agents.<\/p>\n
The Risk of a Machine-to-Machine Web<\/p>\n
There is another danger. AI-generated content is becoming abundant, while AI agents are increasingly consuming information produced by other AI systems. This could create a cycle in which machines generate, summarize and redistribute information with progressively less human involvement.<\/p>\n
The result could be an internet containing enormous quantities of material but with weaker connections to original sources. Provenance may become as important as content itself. Users may need to know not only what an AI system tells them, but where the information originated and how many times it has been transformed.<\/p>\n
The End \u2014 or a Transformation?<\/p>\n
The human internet is not disappearing. Humans will continue to create, communicate and consume information online. But its architecture is changing rapidly.<\/p>\n
The critical question is whether AI agents will remain assistants operating inside a human-centred internet, or whether the web itself will increasingly be redesigned around machine users.<\/p>\n
Cloudflare has already described the emerging agentic internet as one that must be readable, discoverable, callable and payable. That description points toward a fundamental change in how websites, information and commerce may function.<\/p>\n
The greatest transformation of the internet may therefore not be that AI creates more content. It may be that AI becomes the primary intermediary between humans and the web.<\/p>\n
If that happens, the internet will still belong to humans in principle. But increasingly, machines may be the ones doing the browsing.<\/p>\n
<\/p>\n
About Altaf Moti <\/p>\n
\n Altaf Moti is a journalist, columnist, and geopolitical analyst specializing in international security and global finance. As a prolific contributor to various international media platforms, he provides insights into the shifting dynamics of the Middle East and South Asia. With a command of English, Urdu, and Arabic, Moti bridges the gap between regional narratives and global strategic discourse. His work explores the intersection of diplomacy, intelligence, and the evolving world order. <\/p>\n