LinkedIn has been training its AI tool with user data without asking for permission

LinkedIn, a Microsoft-owned company, announced this week in a blog post that "we are updating our User Agreement and clarifying some practices covered by our Privacy Policy" to clarify how it uses personal data to train its artificial intelligence-powered tools, which can generate writing suggestions and post recommendations.

The clarification comes after several media outlets reported that LinkedIn has taken users' personal data to train its artificial intelligence models without properly notifying them. People's accounts were included in the system automatically, which means that their personal information has already been used.

The system was as follows: When members use the professional networking platform, it collects data about their activity, such as their posts, language preferences, login frequency and any comments they may provide. LinkedIn claims the information is used to "fine-tune" its artificial intelligence products.

Independent tech publication 404 Media claimed that "LinkedIn is training its artificial intelligence with data from its users before it updates its terms of service." 

USA Today published comments from LinkedIn spokesman Greg Snapper, who said that "we have always been clear in our terms of service" and emphasized that users have choices regarding the use of their data. Among the data LinkedIn can use to train its A.I. is articles posted by users. In this regard, USA Today claims that if a user posts an article and also mentions certain people's names, LinkedIn's generative writing suggestions feature can include those names. The user can then edit or revise the post before publishing it.

According to LinkedIn's site, opting out prevents LinkedIn and Microsoft from using personal data and content to train models in the future but does not undo or affect training that has already taken place.