A US appeals court has lifted the ban on Perplexity‘s AI shopping agents on Amazon.
The legal logic
The 9th US Circuit Court of Appeals in San Francisco ruled that Amazon was unlikely to succeed in its claim that Perplexity violated federal computer fraud law. The decision rests on who actually accessed the system. The court found that users, not the AI company, accessed Amazon’s platform through the agents.
Under the relevant statute, accessing computers or obtaining information without permission is prohibited. The appeals court sided with Perplexity because the agents acted on behalf of the people logging in. The startup argued that users chose to use the tools, meaning the access was authorised by the individuals, not the software provider.
First major ruling on agentic systems
This is the first time a US federal appeals court has addressed whether AI agents can lawfully access online platforms on behalf of users. Agentic systems plan, reason, and complete tasks with limited human oversight. They browse websites, shop, and process transactions across the web.
The outcome matters for developers building these tools. If the court had upheld the injunction, it would have set a precedent requiring explicit platform permission for every automated action. By overturning it, the court suggests that user consent may be enough to shield such tools from fraud charges.
Amazon’s response and the wider context
Amazon is not accepting the loss. A spokesperson stated, “We respectfully disagree with today’s decision on the preliminary injunction,” adding that the company remains confident and is weighing its options. “We remain confident in our case and are evaluating our next steps.”
Perplexity spokesperson Jesse Dwyer said, “Perplexity will continue to fight for the right of internet users to choose whatever AI they want.”
From the November lawsuit to a reversal on appeal
Amazon sued Perplexity in November, accusing the startup of covertly accessing private customer accounts through its Comet browser and the integrated AI agent. The agent can log into shopping profiles and place orders. Amazon’s complaint alleged security risks and claimed Perplexity had ignored repeated demands to stop. The startup called the lawsuit baseless and a “bald attempt” to keep Amazon customers away from Comet because AI agents “don’t have eyeballs to see the pervasive advertising Amazon bombards its users with.”
In March, a federal court in California temporarily barred Perplexity from using agentic AI on Amazon. Judge Maxine Chesney found “strong evidence” that Perplexity accessed password-protected accounts with users’ permission but without Amazon’s authorization. The startup was also ordered to delete copies of Amazon data. That injunction is now void, but the underlying case hasn’t been decided.
Amazon had argued that third-party AI must operate transparently and respect established APIs, pointing to its own tools like “Buy For Me” and the “Rufus” assistant. Perplexity CEO Aravind Srinivas called it “bullying” and argued that agents acting on behalf of users should have the same rights as human users. The business ties between the two companies add another layer to the dispute. Perplexity is an AWS customer with commitments in the hundreds of millions of dollars, and Jeff Bezos is among the startup’s investors.
