A federal judge just drew a hard line on AI shopping agents. U.S. District Judge Maxine Chesney granted Amazon a preliminary injunction on March 10, blocking Perplexity's Comet browser agent from accessing Amazon's platform to make purchases on behalf of users.
Comet is Perplexity's autonomous shopping tool - a browser agent that logs into your Amazon account, finds products, and completes purchases for you. The problem, according to Amazon, is that while users gave Comet permission to access their accounts, Amazon never authorized a bot to do so. That distinction matters. Judge Chesney found "strong evidence" that Perplexity violated the federal Computer Fraud and Abuse Act (CFAA), the same law used to prosecute hackers, along with a California computer fraud statute.
The Order
The injunction requires Perplexity to stop using Comet to access password-protected parts of Amazon, including Prime subscriber accounts. It also orders Perplexity to destroy any Amazon data it previously collected. The order is paused for one week to give Perplexity time to appeal.
Amazon originally filed its complaint back on November 4, 2025, making this the first major legal action by a major retailer against an AI shopping agent.
A Precedent for Every AI Agent
This ruling goes well beyond Amazon and Perplexity. The core legal question - can an AI agent access a platform on a user's behalf if the platform hasn't explicitly allowed bots? - applies to every agentic AI tool trying to interact with existing websites. If this reasoning holds up on appeal, any AI agent that logs into a third-party service could face similar legal challenges, even when the human user fully consents.
For now, the practical takeaway is clear: user permission alone is not enough. Platforms control who - and what - gets access to their systems. AI companies building autonomous agents will need to negotiate access directly with platforms, or find their products on the wrong side of computer fraud laws.