Don’t miss our latest stories. Add PC Mag As a preferred source on Google.
Update: Now Amazon has filed a nuisance lawsuit, Reuters reports, and has escalated the dispute from its initial cease-and-desist letter.
Amazon says the problem continues to be that it continues to run its Comet AI browser on the site without Amazon’s permission, despite repeated requests for it to stop. AI Agent is allegedly accessing customer accounts without the consent of the company or its customers.
“The corruption of harassment must end,” Amazon says. “The harassment is not allowed to go where it is clearly told it cannot.” That problem corruption involves a code rather than a lockpick.
The case, which Amazon filed in the U.S. District Court for the Northern District of California, also alleges interference with its business by harassing the shopping experience it has built over decades. .
Original Story (11/4):
In a case of big tech versus budding AI startups, Amazon is demanding trouble stop its agent web browser, Comet, from buying products on behalf of humans.
“We have repeatedly requested that Amazing Amazon remove the Comet experience, particularly in light of the significantly degraded shopping and customer service experience it provided,” Amazon says.
As Bloomberg reports, Amazon has been accused of violating its terms of service and committing computer fraud when AI is shopping for a customer. It reportedly asked Plexity to stop using its AI agents on Amazon a year ago, and Plexity agreed. However, that changed with the launch of the Comet.
“We think it’s pretty straightforward that third-party applications that offer to buy from other businesses on behalf of consumers should operate openly and respect the service provider’s decisions whether or not to,” Amazon says.
In our tests, the Comet browser was able to buy a product on Amazon when we asked it to, without requiring us to log in or enter credit card information. He already had access to the account and completed the process independently.
Get our best stories!
Your daily dose of our top tech news
By clicking Sign Up, you confirm that you are 16+ years of age and agree to our Terms of Use and Privacy Policy.
Thanks for signing up!
Your membership has been confirmed. Watch your inbox!
“It took about 30 seconds before he got me to confirm, and he placed the order using my default payment method and address,” says PCMag’s Ruben Cerelli. “It’s definitely easier, and it seems to work, at least on Amazon.”
Anxiety published a strongly worded statement on what it called Amazon’s “aggressive legal threat,” accusing the tech giant of bullying and using litigation to stifle innovation.
“Amazon must love it,” Worry says. “Easier shopping means more transactions and happier customers. But Amazon doesn’t care. They’re more interested in serving your ads, sponsored results, and influencing your purchase decisions with appeals and confusing offers.”
Amazon can take this issue with its customers using another company’s tools, such as the “Help Me Decide” AI shopping assistant it debuted last month to make purchases. The problematic experience also requires customers not to visit Amazon’s website, potentially devaluing and limiting browsing and additional purchasing opportunities.
Recommended by our editors
The concern claims that Amazon is trying to “make life worse,” and argues that the e-commerce giant “shouldn’t forget that it likes to be passionate about products our size and changing the world.”
Still, agentic AI browsers are nascent technologies with known problems. Even Openai admits that its ChatGPT Atlas is flawed and may lead users to buy the wrong products.
“The ChatGPT agent is powerful and helpful, and designed to be secure, but it can still (sometimes surprisingly) make mistakes, like trying to buy the wrong product or forgetting to check with you before taking an important action,” said Dan Stuckey, chief information security officer at Openai. written on x Shortly after the release of Atlas.
Stokey says the main reason would be this Instant injection attackor when a hacker “embeds malicious instructions into websites, emails, or other sources, to try to make the (AI) agent behave in unwanted ways.” In other words, a customer can ask the AI ​​to buy toilet paper, but an injection prompt can instruct the AI ​​to ignore that instruction and buy something else instead.
The response to anxiety does not acknowledge these risks. Instead, it taps into the long-term potential for AI. “With the rise of agentic AI, software is also becoming a labor force: an assistant, an employee, an agent,” he says. “Today, Amazon announced that it no longer believes in your right to have your labor hired, assistant or employee work on your behalf.”
Amazon’s response to AI-powered shopping differs from Walmart, which signed a deal with Openai last month to allow ChatGPT users to shop its catalog. With Openai’s instant checkout technology, shoppers can purchase Walmart items without ever visiting the chat interface or Walmart.com. “It’s virtually agentic commerce,” Walmart said.
Disclosure: Ziff Davis, the parent company of PCMAG, filed a lawsuit against Openei in April 2025, alleging that it infringed Ziff Davis copyrights in training and operating its AI system.
About our expert
Emily Forlini
Senior reporter
Experience
As a news and features writer at PCMag, I cover the biggest tech trends that shape the way we live and work. I specialize in on-the-ground reporting, uncovering the stories of people who are at the center of change—whether it’s the CEO of a high-value startup or the everyday person driving Big Tech. I also cover daily tech news and breaking stories, putting them into context so you get the whole picture.
I came to journalism from a previous career working at Big Tech on the West Coast. This experience gave me an intimate view of how software works and how business strategies change over time. Now that I’m in my master’s in journalism from Northwestern University, I report on my insider knowledge and chops to help answer the big question: Where is it all going?
Read full bio
