The internet was built for human users, but the explosion of —automated programs that perform tasks on behalf of AI systems—is forcing websites to adapt. Until recently, sites could rely on simple bot rules to manage automated traffic. Now, millions of AI agents are attempting to access web content simultaneously, and websites must decide whether to optimize their infrastructure and design for these new visitors.
If sites redesign themselves for AI agents, the human experience could suffer: faster-loading pages, simplified layouts, and machine-readable content might make websites less engaging or useful for people. The tension reflects a deeper question about who the internet is for as AI becomes a primary consumer of web data.
This shift has implications for AI development itself. Better web infrastructure for agents could accelerate AI training and deployment, but at the cost of fragmenting the internet into separate experiences optimized for humans versus machines.
This is an AI-written summary. The original reporting belongs to MIT Technology Review. How we write these · Spot a mistake?