Navigation
Recherche
|
Google tricks ChatGPT into exposing itself
mercredi 29 novembre 2023, 19:27 , par BoingBoing
A new paper shares research from Google DeepMind, the University of Washington, Cornell, Carnegie Mellon University, the University of California Berkeley, and ETH Zurich demonstrating that OpenAI's ChatGPT does memorize immense amounts of data they do not own and personally identifiable information that it should not. — Read the rest
https://boingboing.net/2023/11/29/google-tricks-chatgpt-into-exposing-itself.html
Voir aussi |
56 sources (32 en français)
Date Actuelle
sam. 18 mai - 17:26 CEST
|