Perplexity AI under scrutiny over illegal web scraping

[ad_1]

FILE PHOTO: AI-powered search portal Perplexity is under the scanner for allegedly plagiarising directly from several news outlets and platforms.

FILE PHOTO: AI-powered search portal Perplexity is under the scanner for allegedly plagiarising directly from several news outlets and platforms.
| Photo Credit: The Hindu

AI-powered search portal Perplexity is under the scanner for allegedly plagiarising directly from several news outlets and platforms. Its newly launched feature, Perplexity Pages, which helps people curate content on a specific subject was found copying and pasting stories from multiple news publications, per media reports. 

A Forbes report shared that while these posts had already received tens of thousands of views, they did not mention the publications aside from tiny logos that were easily missed that link out to them. Additionally, the platform had managed to scrape through an article that under the publication’s paywall. 

CEO Aravind Srinivas responded to the allegations clarifying that his platform had been using third-party scrapers that ignored the robots.txt code which explicitly asked web crawlers to not scrape the page. 

A week ago, Srinivas also appeared on Lex Friedman’s podcast and shared that he had built a tool that could scrape through X by pretending to be an academic researcher for API access. 

(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)

Musk, who has been against AI companies scraping X data to train their AI models, updated the social media platform’s policies to ban crawling and scraping and cutting off free access through the API. 

A report published by Wired also revealed that Amazon’s cloud division, AWS has launched an investigation into Perplexity AI to determine whether the startup was violating rules by scraping websites despite attempts to prevent it. 

An Amazon spokesperson shared while the protocol wasn’t legally binding, the terms of service were and scrapers normally respected it. All customers of the company’s cloud division must comply with the robots.txt guidelines during website crawls, the source noted. 

Srinivas’ startup is hot property at the moment. 

According to reports, Japanese tech investor SoftBank Group is said to be investing between $10 million and $20 million in the startup bringing its valuation to $3 billion.

[ad_2]

Source link

Leave a Comment