Websites Are Closing Their Doors to AI Agents

OpenAI, Meta, and Google want AI agents to do everything for you. But popular services on the web aren't letting that happen.

By Emma Roth

The major AI platforms have spent years promising a future in which autonomous agents book your travel, file your taxes, and shop on your behalf. By 2026, that vision is colliding with a hard reality: many of the web's most popular services are actively shutting agents out.

The Promise vs. the Blockade

OpenAI, Meta, and Google have each rolled out agentic features designed to operate websites on a user's behalf—clicking, filling forms, and completing transactions. But a growing number of high-traffic destinations, from e-commerce platforms to financial services portals, are deploying bot-detection systems, blocking user-agent strings, and updating terms of service to prohibit automated access.

Why Sites Are Pushing Back

Site operators cite a familiar set of concerns: server load from automated traffic, fraud and account-takeover risks, liability when an agent acts without clear human oversight, and the loss of control over user experience and monetization. In 2026, the dispute has sharpened as AI crawlers and agents increasingly blur the line between legitimate automation and unwanted scraping.

The Agentic Web Hits a Wall

The result is a fragmented landscape. Some sites are experimenting with official agent protocols and verified bot partnerships; others are locking down entirely. Until standards for authentication, consent, and compensation mature, the dream of a fully agent-driven web remains uneven—and contested.

via The Verge AI