LinkedIn came under fire earlier this year for allegedly using private messages of its Premium members to train its AI models. Now, the Microsoft-owned company is updating its terms of service (ToS) to explicitly detail the new ways in which customer data stored on LinkedIn may be utilized.
In an email sent out to users and seen by Neowin, LinkedIn has informed customers that from November 3, 2025, data belonging to users in the European Union (EU), European Economic Area (EEA), Switzerland, Canada, and Hong Kong will be processed for training AI models that are used to generate content. The company says that this will enhance your LinkedIn experience and connect you to more relevant opportunities on the platform.
The data used to train models will include your profile and any public content that you post on LinkedIn. Fortunately, LinkedIn will not access your private messages. Another thing to keep in mind is that this data will be used by all LinkedIn affiliates, which include LinkedIn Corp., LinkedIn Ireland, and Microsoft. Affiliates will be able to use your data by default unless you toggle this capability off through the dedicated setting here.
Moreover, users in the U.S. will be sharing their LinkedIn data with Microsoft by default for targeted advertising. LinkedIn says:
LinkedIn may use information that we receive from others about your engagement with their sites and services, in order to improve our tools for targeting ads. While we do not use your off-LinkedIn engagement data to predict your interests at an individual level, we do use it to improve our tools for targeting ads overall. Many U.S. states consider ads selected based on an individual’s activities outside of a company’s own websites or applications to be “targeted advertising” and provide you a right to opt out of having your data used for this purpose.
The option to disable data sharing with Microsoft can be found here while the ability for LinkedIn to use data from other sites for targeted advertising can be toggled off here.
Other updates to the ToS clarify some terms and violations such as the use of deep fakes and impersonation. You can find out more details here, before they officially come into effect on November 3, 2025.