A federal judge in the US has determined that Perplexity AI’s AI shopping agent should no longer be allowed to access Amazon’s platform. According to the judge, Amazon had strong indications that the AI tool may be violating a US anti-hacking law by accessing Amazon users’ accounts without permission.
The tussle between Perplexity and Amazon dates back to November 2025, when the e-commerce giant filed a lawsuit against the answer engine developer. Amazon accused the company of automatically shopping on behalf of users via its Comet browser, even after Amazon implemented technical blocking and sent a cease-and-desist letter.
Amazon argued that in doing so, Perplexity may have violated the Computer Fraud and Abuse Act (1986) and a California anti-hacking law. Those laws prohibit unauthorised access to computer systems.
In addition, Amazon claimed that the AI agent does not clearly identify itself as AI software and that the Comet browser is set up so that activities appear to come from Google Chrome, and a human user.
Amazon said other AI companies had transparency about their bots and respected access rules.
Perplexity disputed the allegations. The company said that as users themselves give permission for the AI agent to shop for them, Amazon was trying to prevent customers from using alternative tools.
According to Perplexity, Amazon’s motive is mainly commercial. AI agents do not see ads and cannot be enticed to make additional purchases, while ads are an important part of Amazon’s shopping experience.
The ruling is preliminary (an injunction) and an appeal has yet to be heard. Still, the case could become important, as it could set a precedent for who has control over AI agents acting on behalf of users on websites.
Emerce
Ltd
