Judge halts AI shopping tool
A federal judge has temporarily stopped Perplexity from using its AI agent to shop on Amazon. The ruling came down on Monday, with Judge Maxine Chesney granting Amazon’s request for a preliminary injunction. She found that Perplexity likely broke computer fraud laws by accessing password-protected Amazon accounts without permission.
Perplexity now has one week to appeal before the injunction takes full effect. The company must pause its Comet AI browser’s shopping capabilities and destroy any Amazon data it collected through this activity. The tool won’t be able to enter password-protected sections of Amazon while the case moves forward.
How the AI operated
Amazon’s lawsuit, filed back in November, claims Perplexity’s Comet browser used an AI agent to perform automated actions within user accounts. The e-commerce giant says the tool pretended to be a human Chrome user while actually running automated shopping processes.
What’s interesting, I think, is how Amazon describes the tool bypassing their safeguards. They implemented technical blocks and sent legal warnings, but according to their complaint, Perplexity kept operating anyway. That’s a pretty bold move if true.
Legal implications
This case touches on some important questions about AI and terms of service. Amazon argues that Perplexity violated its terms of use by using automated agents to access customer accounts. The company is seeking damages and wants a permanent court order preventing Perplexity from using AI agents to access its systems.
Judge Chesney’s decision suggests she sees merit in Amazon’s arguments, at least enough to issue this temporary block. The finding that Perplexity likely violated both federal and California computer fraud laws is significant. These laws typically deal with unauthorized access to computer systems.
What happens next
Now we wait to see if Perplexity appeals within that one-week window. If they don’t, the injunction takes effect and their AI shopping tool stays offline for Amazon accounts. The broader case will continue through the legal system, which could take months or even years to resolve completely.
This situation raises questions about how AI companies should approach integration with existing platforms. There’s a line between innovation and violation of terms, and this case might help define where that line sits for AI agents. Amazon clearly views this as crossing into unauthorized territory.
It’s worth watching how this develops, because similar issues could arise with other AI tools trying to interact with established platforms. The outcome might set some precedents for what’s acceptable in terms of AI automation on third-party services.
