Home SECURITY Google warns against hallucinating chatbots

Google warns against hallucinating chatbots

0
Google warns against hallucinating chatbots

[ad_1]

Google warns against hallucinating chatbots

AI bots provide answers that look credible, but are not based in any way on facts.

Senior Vice President Google and Head of Google Search, Prabhakar Raghavan, warns users about the importance of reviewing information received from AI-based chatbots.

The variety of AI we are talking about sometimes leads to the creation of fictional responses, which we call hallucinations, Mr. Raghavan said. He also stated that one of the important tasks is to minimize such cases.

When hallucinating, AI bots provide answers that look authentic but are not based in any way on facts. Developers should minimize these disruptions to AI bots.

This week, security researchers at Check Point Research declared that hackers have developed a way to bypass ChatGPT’s restrictions and are using it to sell services that allow cybercriminals to create malware and phishing emails.

It also became known that OpenAI collects a wide range of other user information : user IP addresses; browser type, its settings; data about user interaction with the site; the type of content they interact with; the features they use; actions they take. All this information can be freely collected by OpenAI in accordance with the company’s privacy policy.

[ad_2]

Source link

www.securitylab.ru

LEAVE A REPLY

Please enter your comment!
Please enter your name here