OpenAI launches web crawler ‘GPTBot’ amid plans for next model: GPT-5

Synthetic intelligence agency OpenAI has launched “GPTBot” — its new internet crawling device which it says may doubtlessly be used to enhance future ChatGPT fashions.

“Net pages crawled with the GPTBot consumer agent could doubtlessly be used to enhance future fashions,” OpenAI stated in a brand new weblog submit, including it may enhance accuracy and develop the capabilities of future iterations.

An internet crawler, generally referred to as an internet spider, is a sort of bot that indexes the content material of internet sites throughout the web. Search engines like google like Google and Bing use them to ensure that the web sites to point out up in search outcomes.

OpenAI said the online crawler will acquire publicly out there information from the world broad internet, however will filter out sources that require paywalled content material, or is understood to collect personally identifiable data, or has textual content that violates its insurance policies.

Breaking

OpenAI simply launched GPTBot, an internet crawler designed to robotically scrape information from your entire web.

This information will likely be used to coach future AI fashions like GPT-4 and GPT-5!

GPTBot ensures that sources violating privateness and people behind paywalls are excluded. pic.twitter.com/oR3kY4buaU

— Shubham Saboo (@Saboo_Shubham_) August 7, 2023

It needs to be famous that web site house owners can deny the online crawler by including a “disallow” command to a normal file on the server.

The brand new crawler comes three weeks after the agency filed a trademark software for “GPT-5,” the anticipated successor to the present GPT-4 mannequin.

The applying was filed at america Patent and Trademark Workplace on July 18, and covers the usage of the time period “GPT-5,” which incorporates software program for AI-based human speech and textual content, changing audio into textual content and voice and speech recognition.

OpenAI has filed a trademark software for:

“GPT-5”

which incorporates “software program for”:

“the synthetic manufacturing of human speech and textual content”

“conversion of audio information information into textual content”

“voice and speech recognition”

“machine-learning primarily based language and speech processing”

pic.twitter.com/54aJBovDNB

— YK aka CS Dojo (@ykdojo) August 1, 2023

Nevertheless, observers could not wish to maintain their breath for the subsequent iteration of ChatGPT simply but. In June, OpenAI’s founder and CEO Sam Altman stated the agency is “nowhere shut” to starting coaching GPT-5, explaining that a number of security audits should be performed previous to beginning.

In the meantime, Issues have been raised over OpenAI’s data-collecting ways of late, notably revolving round copyright and consent.

Japan’s privateness watchdog issued a warning to OpenAI about accumulating delicate information with out permission in June, whereas Italy briefly banned the usage of ChatGPT after alleging it breached numerous European Union privateness legal guidelines in April.

In late June, a category motion was filed towards OpenAI by 16 plaintiffs alleging the AI agency to have accessed personal data from ChatGPT consumer interactions.

If these allegations are confirmed to be correct, OpenAI — and Microsoft, who was named as a defendant — will likely be in breach of the Laptop Fraud and Abuse Act, a legislation with a precedent for web-scraping circumstances.