Major Websites and Retailers Move to Block Autonomous AI Agents
Top platforms like Amazon, Yelp, and major banks are restricting AI agents from taking actions online, forcing developers to rethink how autonomous bots interact with the web.
Major web platforms and enterprises are increasingly restricting autonomous AI agents from navigating their websites and executing actions on behalf of users. While technology giants like OpenAI, Meta, Google, xAI, and Instinct promise that AI agents will automate routine tasks—such as booking restaurant tables, clearing inboxes, and buying merchandise—the real-world web is proving resistant. Recent reports indicate widespread blockades affecting toolsets like Meta's Muse, xAI's Grok, OpenAI's Dots, and Instinct's agent, creating significant hurdles for the consumer adoption of autonomous digital assistants.
A growing wall of resistance across industries
Companies across multiple sectors are actively shutting down agentic web activity. Review platform Yelp announced that it restricts AI agents unless they subscribe to its paid data licensing program, while Delta Air Lines prevents autonomous bots from purchasing flight tickets. Financial institutions including JPMorgan Chase and Charles Schwab confirmed blocking AI agents over security considerations. Meanwhile, users report that platforms previously accessible to their agents, such as Cars.com and Amtrak, have quietly ceased working.
E-commerce giant Amazon has been at the forefront of this movement. Last year, Amazon banned Perplexity's agentic shopping functionality and pursued legal action against the startup, stating that third-party companies buying items for customers must operate transparently and respect provider decisions. Perplexity countered by alleging that Amazon fears losing human engagement on its advertising displays, asserting that an AI assistant functions as an extension of the user. Although a court initially ordered Perplexity to halt its shopping agent on Amazon, an appeals court overturned the ruling in August, leading Amazon to file an amended complaint. Similar friction persists today, with Amazon declaring that continued site access by Meta's Muse violates its terms of service, alongside reported access issues for Grok, Dots, and Instinct. In response, xAI owner Elon Musk commented that he might consider launching a competitor to Amazon.
Anti-bot infrastructure and strategic partnerships
Beyond explicit bans, generic anti-bot protections and security frameworks are tripping up AI agents. Cloudflare introduced a one-click block for AI crawlers in 2025 and subsequently began blocking crawlers that index web content for search and training simultaneously. Furthermore, Reddit users report that Google has suspended accounts using Meta's Muse due to flagged suspicious activity, while others encountered errors attempting to check out on Walmart despite Meta's formal partnership with the retailer. Chris Canal, CEO of AI safety evaluator EquiStamp, pointed out that existing security protections designed to prevent hacking or scamming naturally trigger against AI agents. Canal warned that these barriers could ultimately make web navigation more tedious, forcing human users through additional verification steps like CAPTCHAs even on everyday retail sites.
To bypass direct conflicts, some AI developers are seeking formal commercial agreements. Meta stated in a recent post that rejecting a user's personal agent equates to turning away the consumer, noting it prefers collaborating with businesses. Meta has formed partnerships with retailers such as Gap, Sephora, Best Buy, and Walmart to enable Muse to shop on their storefronts, while e-commerce platform Shopify has similarly opened its merchant network to AI agents.
What it means for developers
For software engineers building on agentic workflows, the rising tide of web blocks highlights a shift away from raw web scraping and browser automation toward official APIs and structured integrations. When building applications powered by LLMs, developers can try top AI models cheaply through one API at https://apixoai.online, but relying solely on autonomous web agents to execute front-end interactions remains risky due to changing site policies and active anti-bot systems. As Chris Canal from EquiStamp noted, active efforts by web hosts to block agents directly diminish the practical utility that frontier model developers claim these systems possess.
Developers must weigh whether to build agents that rely on direct web interaction or focus on ecosystem partnerships. With major entities like OpenAI and Google facing undocumented roadblocks, and Instinct declining to comment on current agent limitations, software creators will likely need to rely more heavily on enterprise data partnerships, licensing agreements, and standardized API protocols to ensure long-term reliability for autonomous tasks.
Source: Websites are closing their doors to AI agents — The Verge AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
