End of the Click

For nearly three decades, the internet thrived on a simple economic bargain: users got free content, and publishers earned revenue through ads, subscriptions, or affiliate links. But that equilibrium is now collapsing. Artificial intelligence ~ ironically built on the bones of that same free content ~ is disrupting the very web it learned from. Chatbots powered by large language models are fast replacing the need to visit web sites altogether. Whether it’s a travel itinerary, a product review, or a medical explanation, AI can now synthesise information instantly, often more fluently than the original source. The result is a sudden and severe drop in web traffic to traditional publishers, educational institutions, forums, and independent creators. And with that decline goes the core incentive to produce and maintain quality information on the open web.

Already, niche blogs, independent science sites, and community forums are disappearing ~ starved of visitors, stripped of purpose, and buried beneath AI-generated summaries that rarely link back. At the heart of this shift is the decoupling of information from its origin. While search engines at least directed users toward a source ~ offering visibility, branding, and monetisation ~ AI responses are often unattributed. In many cases, the model has been trained on scraped data without permission, leaving publishers uncompensated even as their intellectual labour becomes part of AI’s output. This raises not only ethical and legal concerns but also structural ones: who will keep producing content when distribution and revenue are siphoned away?

Advertisement

The responses so far have been fragmented. Some content providers are attempting to wall off their data, implementing bot-blockers or demanding licensing fees. Others are litigating, betting on courts to enforce copyright norms in a context that the law was never designed for. Meanwhile, a few have chosen to strike deals with AI firms ~ trading exclusivity for survival. But all these are temporary measures, not long-term solutions. There is still time to rethink the architecture of the digital commons. One way forward is a licensing ecosystem where AI companies compensate content creators, perhaps using micropayments or shared-revenue models.

Another is embedding attribution within AI outputs, so users can choose to visit the original source. More radically, regulatory intervention might be required to compel transparency, enforce fair use boundaries, or even slow the extractive pace of AI development. Ultimately, this is not just about the web economy ~ it is about information integrity, public trust, and the sustainability of human knowledge systems. If AI becomes the primary interface for consuming information, then it must also become a stakeholder in preserving the conditions that make reliable information possible. The open web may not survive in its current form. But what replaces it must not be a black box. Without reinvention, we risk a future where the web is no longer a living network of ideas ~ but a fossil record mined by machines.