Amazon Reviewing Whether Perplexity AI Incorrectly Scraped Online Content

Amazon is reviewing allegations that artificial intelligence startup Perplexity AI is scraping content — including from prominent news sites — without consent.

Amazon spokeswoman Samantha Mayowa confirmed Friday that the tech giant was evaluating information it received from news agency WIRED, which published an investigation earlier this month that said Perplexity appeared to be scraping content from websites that had blocked access to such practices. Perplexity uses Amazon Web Services, also known as AWS, servers.

“Amazon’s Terms of Service prohibit abusive and illegal activities, and our customers are responsible for adhering to these terms,” Mayowa said in a prepared statement. “We routinely receive reports of alleged abuse from a variety of sources and work to engage our customers to understand those reports.”

Perplexity spokeswoman Sarah Blatnick said Friday that the company has determined that Perplexity-controlled services do not crawl websites in any way that violates AWS’s terms of service.

The San Francisco-based AI search startup has been a darling of high-profile technology investors, including major investors like Amazon founder Jeff Bezos. But in the past few weeks, the company has found itself in trouble amid accusations of intellectual theft.

Aravind Srinivas, CEO of Perplexity, offered a strong defense of the startup after it published a brief news story with information and wording similar to a Forbes investigative story. She did so without citing the media or asking for their permission. Forbes later said it had found similar “copycat” stories taken from other publications.

Separately, The Associated Press found another Perplexity product. I invented fake quotes. From real people.

See also  Intel Stock Up as $11 Billion Apollo Deal Closes to Completion: Wall Street Journal

Srinivas said in an interview with The Associated Press earlier this month that his company “did not steal content from anyone. Our engine does not train on anyone else’s content,” in part because the company simply collects what other companies’ AI systems generate. .

But he added: “Forbes has accurately indicated that they would prefer to highlight the source more clearly.” Sources are now being highlighted more prominently, he said.

___

Associated Press reporters Matt O’Brien and Sarah Parvini contributed to this report.

Leave a Reply

Your email address will not be published. Required fields are marked *