As of November 3, 2025, LinkedIn will begin using certain personal data of its members to train its generative artificial intelligence models. The announcement was made on September 18 by the platform in an official post from Blake Lawit, its Chief Global Affairs & Legal Officer, followed by a user notice informing about an update to the Terms of Use.
The information leveraged to train the platform’s AI includes profile data (name, photo, career history and education, skill endorsements, location), public messages and replies to posts, as well as contributions to groups and certain interactions with recruiters. The questions posed to LinkedIn’s AI are also taken into account.
By contrast, private messages and salary information will not be part of the data collected. LinkedIn also notes that when there is any doubt about a user’s age—such as when a user is enrolled in a secondary school—those data will not be used, even if the sharing option is enabled.
Private messages and salary information will not be collected
Members have an option to refuse the use of their data for training purposes. This preference can be changed within account settings, via the dedicated section titled “Use my data to train LinkedIn’s generative AI creation models.”
In September 2024, following a query from the ICO, the British regulator, LinkedIn paused the use of member data in the United Kingdom for training its AI models, after having quietly enabled by default consent from users worldwide for the same data usage. At the same time, the company halted data collection from users in Canada, mainland China, Hong Kong, the European Economic Area, and Switzerland.
“We are pleased that LinkedIn has taken our concerns into account and adjusted its approach to allow users to object more easily to processing and to have a longer window to do so,” commented an ICO spokesperson. “We made it clear that when organizations intend to use user information to train generative AI models, they must implement effective safeguards before proceeding with training, notably by providing a clear and straightforward way for users to opt out.”
