“Dead Internet” Theory Becomes Reality: Bot Traffic Surpasses Humans

The “dead Internet” theory, long debated in the digital world, is no longer just a hypothesis—it has become a reality confirmed by numbers. According to recent data, the majority of total traffic on the global network is now generated not by humans, but by automated programs, or bots. This situation is expected to pose a serious challenge for the digital economy and security systems. This is reported by Ixbt.com reports .

According to a report by Cloudflare, by the end of June, nearly 58 percent of all requests made to web pages came from bots. For the first time in Internet history, this figure has surpassed human activity. Previously, Cloudflare CEO Matthew Prince predicted that such a scenario would only happen by 2027, but the rapid development of AI technologies has accelerated this process by several years.

The surge of AI agents

Experts from Human Security noted that over the past year, the activity of AI agents capable of independently following links, filling out forms, and acting on behalf of users has increased by 8,000 percent. A Thales report previously projected that bot traffic would reach 53 percent by 2025, but in reality, these figures have already been surpassed. Worst of all, nearly 40 percent of this automated activity is malicious in nature.

This trend is clearly visible on major financial and tech platforms. For example, 70 percent of requests to the Stripe payment system's API are made by software agents. On the Alpaca platform, the share of bots jumped from 10 percent to 30 percent in just one quarter. This indicates a fundamental shift in how the Internet operates.

Traditional defense systems are becoming obsolete

Today, traditional security methods like CAPTCHA are increasingly failing to distinguish modern AI agents from humans. Therefore, Cloudflare introduced Precursor, a system that analyzes the entire session rather than individual user actions. Experts believe that current systems mistakenly classify 7-15 percent of regular users as bots, while many smart bots easily slip past the filters.

To regulate the situation, giants like Google, Microsoft, Mozilla, and Cloudflare have started working on a new protocol called PACT. This technology allows verifying that a visitor is indeed human without revealing their identity. This is a life-or-death issue for the current Internet ecosystem, which is based on an advertising model and analyzes real user behavior.

This process is expected to expand further in the future. According to PitchBook analysis, AI agents currently cover only 1 percent of the potential workload they are capable of performing. This means that the human share on the Internet will decrease even further in the coming years. Given this, Cloudflare announced that starting in September, data-harvesting AI crawlers must pay publishers for using their content.