The bifurcation of the internet is coming | Tech Radar

Overview

News, deals, reviews, guides and more on the newest smartphones

News, deals, reviews, guides and more on the newest computing gadgets

Details

Start exploring exclusive deals, expert advice and more

Unlock and manage exclusive Techradar member rewards.

Agents dominate web, splitting human and machine internet

When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

Unlock instant access to exclusive member features.

Get full access to premium articles, exclusive features and a growing list of member rewards.

We’ve quietly crossed a threshold. For the first time in the internet’s history, bots outnumber humans online. This milestone signals a deeper shift in how the web functions and for whom (or what) it’s designed for. What started as an ecosystem built by and for humans is increasingly becoming one optimized for agents.

There are 8 billion humans today, and there will likely be about 8 billion in a decade. Human growth is linear. Agent growth won’t be. Within the next few years, I believe an 80/20 internet will emerge: 80% agentic traffic, 20% human.

AI agents are already crawling, scraping, synthesizing, and increasingly generating content at a scale no human workforce could match—reshaping the web in real time. Within ten years, we may see hundreds of billions—perhaps even close to a trillion—agents operating online.

Tame your AI gremlins before the chaos becomes permanent

AI bot web traffic is closing in on human usage, experts warn

AI and bots now account for more internet traffic than humans, report finds

The result: a bifurcated web, one layer built for bots, the other for people.

The agentic layer: structured, searchable, and synthetic

The agentic layer of the web is already taking shape. It’s built on structured data, robust metadata, and machine-readable formats that make it easy for AI systems to extract meaning.

Search engines and large language models rely on this layer to train, infer, and generate. The next iteration of autonomous agents will be wholly dependent on it to negotiate, transact, and make decisions.

From a technical standpoint, this layer is incredibly efficient. It favors semantic markup, schema.org compliance, and content that’s optimized for comprehension by agents. But it also comes with tradeoffs. The more we build for agents, the more uniform the digital landscape becomes. Creativity gives way to clarity.

Emotion gives way to precision. The internet that was once chaotic, occasionally quirky, and deeply human starts to flatten under the weight of optimization.

This shift also hints at a deeper architectural evolution. Instead of traditional web pages, agents will increasingly consume APIs, knowledge graphs, data streams, and agent-friendly formats that allow for direct retrieval rather than navigation.

Instead of surfing web pages like humans, AI agents will move through inter-connected knowledge graphs, where each fact links to the next. In this world, the graph becomes the interface AI relies on, and the traditional web page takes a supporting role, serving humans while agents interact directly with structured data.

Five AI agent predictions for 2026: The year enterprises stop waiting and start winning

The pilot phase is over. Here’s what’s next for enterprise AI automation

AI‑first browsers and the end of the pageview economy

The human layer: emotional, experiential, and scarce

In contrast, the human web will evolve into a different kind of space that values experience, emotion, and authenticity.

As bots consume the structured web, people will seek refuge in experiences that feel alive, unpredictable, and real. We’ll see a rise in closed communities, live interactions, and content that resists automation because it’s inherently personal or experiential.

Think of this as the artisanal internet - a counterweight to the algorithmic one. It’s where creators, brands, and organizations will focus less on reach and more on resonance; where human creativity will thrive.

The goal won’t be to feed the agentic layer, but to connect meaningfully within the human one. In many ways, this will look like a return to the internet’s roots: smaller networks, more intentional conversations, and content that doesn’t need to rank to matter.

This bifurcation won’t just be philosophical, but will reshape value exchange online. The agentic internet will become the backbone of automation, powering decisions, transactions, and supply chains. Its value will lie in speed, scale, and interoperability.

The human internet, on the other hand, will trade in trust, context, creativity, and emotional intelligence, qualities agents can’t fully replicate.

This is where the economics get interesting. Value flows will diverge. Subscription models, creator economies, and premium content will likely skew toward human audiences, while data licensing and training-data markets grow around the agentic mode.

Brands will face a fundamental question: if their structured data is scraped to train future models, is that value exchange or uncompensated contribution? We may see new licensing frameworks or data-monetization mechanisms emerge as companies push for compensation within AI training pipelines.

Marketing and SEO will need to evolve, too. Advertising may shift toward agentic engagement, where the “customer” is an agent evaluating structured product detail pages, APIs, or machine-readable offers.

Very few marketers have seriously considered what it means to market to entities with infinite memory, near-zero attention constraints, and superhuman data processing capabilities.

Agents won’t respond to emotional hooks, scarcity tactics, or brand storytelling in the same way humans do. They will evaluate completeness, consistency, price efficiency, interoperability, and verified performance.

Advertising targeting humans will become more experiential, story-driven, and community-centric. As search becomes retrieval- and conversation-driven, traditional rankings will matter less. Companies will optimize for agent interfaces, not just human interfaces.

Duplicate content rules—a pillar of SEO—will need to adapt so brands aren’t penalized for producing standardized, machine-readable data across multiple surfaces. ROI models will split: human engagement measured by depth and trust; agentic engagement measured by precision, accessibility, and machine interpretability.

As AI-generated content continues to flood the web, authenticity will become both more valuable and harder to verify. When bots write for bots, the risk of misinformation compounds exponentially.

Distinguishing between human-originated and machine-generated content will require new forms of digital watermarking, tracking, and verification protocols.

This is where governance and technical stewardship come into play. Just as the early internet needed protocols for security and privacy, the next era will need standards for agentic transparency.

Businesses that lead here won’t just mitigate risk, they’ll invest in ethical AI practices, explainability, and traceability to bolster the next generation of digital trust.

The question for leaders isn’t whether or not this split will happen. It’s how we build responsibly on both sides of it. For startups, that means designing products that are interoperable across human and agentic experiences.

For enterprises, it means rethinking data strategy; not just what information is available, but who or what it’s for. And for policymakers and technologists, it means creating frameworks that ensure the agentic internet enhances rather than erodes the human one.

We’ve reached a familiar moment in the evolution of technology: when innovation density and velocity precipitates a step change. As someone who’s spent decades building at the intersection of human and machine, the pattern is plainly observable, and I’m confident that this change has already started.

The bifurcation of the internet is a reaction to the sheer volume of agentic intelligence now coursing through it.

The opportunity, as always, lies in adaptation. The leaders who thrive in this next era won’t be those who fight automation, but those who design for coexistence.

This article was produced as part of Tech Radar Pro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of Tech Radar Pro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

You must confirm your public display name before commenting

1 Claude Cowork is now available for enterprise use, adds analytics, access controls and more

2 You, Me & Tuscany's unhinged Bridgerton reference was 'absolutely planned' — but Regé-Jean Page still had to 'ignore' it

3 Open AI halts £31 billion Stargate UK project over rising energy costs and regulatory deadlock

4 Some Samsung Galaxy S22 Ultras are being hijacked after a factory reset, with no clear fix

5 Top Word Press Slider plugin hijacked to spread malware — here's what to look out for

Tech Radar is part of Future US Inc, an international media group and leading digital publisher. Visit our corporate site.