LinkedIn suspend l'entraînement de modèles d'IA à partir de données d'utilisateurs

Le commissaire à la protection de la vie privée, Philippe Dufresne, affirme avoir demandé des informations à LinkedIn sur ses pratiques d'entraînement de modèles d'intelligence artificielle. M. Dufresne attend de comparaître devant le Comité permanent de l'accès à l'information, de la protection des renseignements personnels et de l'éthique à Ottawa, le 19 novembre 2024. LA PRESSE CANADIENNE/Patrick Doyle

Canada's privacy commissioner says LinkedIn has paused training its artificial intelligence models on data from ºÚÁϳԹÏÍø users.

Philippe Dufresne says his office was informed of the pause by the tech platform, which he had reached out to after media reports suggested LinkedIn was training AI models with data from ºÚÁϳԹÏÍø users it had not informed of the practice first.

Dufresne says he has requested information from LinkedIn about the company’s training practices as well as how it obtains consent from its members.

He says LinkedIn believed it had taken a "privacy protective" approach to training its AI models but will now work to ensure it has met the requirements of privacy laws before restarting the practice.

Dufresne says even when personal information is publicly accessible, it is subject to privacy laws and must be adequately protected.

LinkedIn suspended using U.K. user data to train its artificial intelligence models in September, after Britain's information commissioner raised similar concerns.

This report by ºÚÁϳԹÏÍø was first published Dec. 10, 2024.

ºÚÁϳԹÏÍø. All rights reserved.

More Politics Stories

Sign Up to Newsletters

Get the latest from ºÚÁϳԹÏÍø News in your inbox. Select the emails you're interested in below.