ChatGPT developer introduced a tool for recognizing text created by AI
[ad_1]
OpenAI, developer of the ChatGPT neural network, announced about launching a classifier capable of recognizing human-written text based on AI-generated material.
The tool will help fight against AI-based disinformation campaigns, malicious use of neural networks in the educational process, and cases where chatbot texts are trying to pass off as human.
The classifier is not flawless, OpenAI emphasized: during tests, it identified 26% of AI-generated text as “probably written by AI”, and 9% of human-written text was marked as AI generated. Moreover, the accuracy of the results increases with the volume of the text being checked, the company noted.
“We have made this classifier publicly available…Our work on AI-generated text detection will continue and we hope to share improved methods in the future,” OpenAI concluded.
In November 2022, OpenAI launched the chatGPT neural network capable of generating texts, answering questions, and writing codes. The chatbot was able to collect the first million free users in just five days. OpenAI was established in 2015 as a foundation for research into artificial intelligence (AI).
[ad_2]
Source link
