Goliath Super Intelligence
IndustryOctober 18, 20261 min read

Websites Push Back Against AI Agents From OpenAI, Meta, and Google

Major online services including Yelp, Delta, JPMorgan Chase, and Amazon are restricting autonomous agents, citing data licensing fees, security concerns, and ad revenue protection.

OpenAI, Meta, Google and newer entrants such as Instinct have been promoting autonomous AI agents that can browse the internet, make reservations, manage inboxes and complete purchases on behalf of users. The promise of a single assistant handling diverse tasks has driven rapid development, but the broader web ecosystem is increasingly resistant to unrestricted access.

Within weeks, users of Meta’s Muse, xAI’s Grok, OpenAI’s Dots and Instinct’s assistant reported that several high-traffic sites stopped responding to their requests. Yelp disclosed that only agents enrolled in its paid data-licensing program may retrieve listings, while Delta Air Lines blocked automated ticket purchases. JPMorgan Chase and Charles Schwab also denied agent access to account data, citing security risks. Reports of sudden loss of service on Cars.com and Amtrak suggest additional, undisclosed blocks.

Amazon has taken a particularly defensive stance, arguing that unrestricted agents would reduce human exposure to its advertising inventory. The e-commerce giant previously sued Perplexity for allowing its bot to shop on the site; a district judge ordered a halt, but an appeals court reversed that decision in August, prompting Amazon to file an amended complaint. Similar access denials have been reported for Muse, Grok, Instinct’s bot and OpenAI’s Dots.

Meta defended its Muse assistant, stating that “turning away a personal agent” would also turn away the underlying consumer. Rather than contest Amazon directly, the company announced collaborations with retailers such as Walmart, Gap, Sephora and Best Buy, and noted that Shopify’s marketplace now permits AI agents to complete purchases. These moves aim to demonstrate commercial viability while addressing merchant concerns.

Web-infrastructure provider Cloudflare has also contributed to the tightening of access. After rolling out a one-click block for AI crawlers in 2025, the firm expanded the feature to target bots that scrape sites for both search-indexing and AI-training data. Cloudflare argues that the surge of autonomous agents strains bandwidth, overloads email systems and jeopardizes fair competition for limited-supply items such as concert tickets and reservation slots.

Sources

  1. Websites are closing their doors to AI agents The Verge

More reports

Industry · October 18, 2026 · 1 min

Anthropic reports fewer junior hires and rising meeting time as AI reshapes work

Anthropic's October 2026 report shows junior technical staff dropped from 21 percent to 14 percent and that meeting time rose as Claude's text output increased.

Industry · October 18, 2026 · 1 min

Replit switches default AI model to OpenAI, citing cost and performance

At TechCrunch Disrupt, Replit CEO Amjad Masad announced OpenAI's ChatGPT will be the platform's default model, while Anthropic remains optional and enterprise revenue is set to exceed half by 2027.

Industry · October 18, 2026 · 2 min

Meta Strips Clause Allowing Unrestricted Scraping of News Content for AI Training

The company removed a term that had let it harvest any website material, including news articles, to feed its artificial-intelligence systems, after criticism from scholars and policy groups.