Updated
Updated · The Verge · Oct 17
Websites Block AI Agents Over Security and Licensing, Undercutting $0-Friction Web Automation
Updated
Updated · The Verge · Oct 17

Websites Block AI Agents Over Security and Licensing, Undercutting $0-Friction Web Automation

3 articles · Updated · The Verge · Oct 17

Summary

  • Yelp, Delta, JPMorgan Chase and Charles Schwab are among companies blocking AI agents from booking, scraping or accessing accounts, limiting tools from Meta, OpenAI, xAI and Instinct.
  • Security risks, paid data-licensing demands and anti-scraping rules are driving the clampdown, while some sites that previously worked with agents—such as Cars.com and Amtrak—now appear inaccessible.
  • Amazon has become a central flashpoint: it warned Meta’s Muse over policy violations, previously sued Perplexity over agentic shopping, and is still pursuing an amended complaint after an August appeals-court reversal.
  • Even where sites do not explicitly ban agents, anti-bot systems are tripping them up; users have reported Google account flags tied to Muse and checkout problems on Walmart despite Meta’s retail partnerships.
  • Cloudflare’s 2025 one-click AI crawler block has reinforced the broader pushback, highlighting fears that AI bots strain infrastructure and could make the web harder for humans to use.

Insights

If websites treat AI agents as security risks instead of user tools, who will control online shopping, booking, and account access next?
Did the Perplexity-Amazon court ruling help users, or is the real future of AI agents hidden in partnerships, APIs, and verified identity?