LinkedIn has started training its AI with public data from European users. Here’s what information is used and how to oppose the automatic sharing of your data.
Twenty-two years of CV, careers and skills. Twenty-two years of posts, comments and recommendations. An immense one archive which tells the work history of hundreds of millions of people: LinkedIn it was officially launched in 2003 and from November 3, 2025 it decided to forage its own artificial intelligence with that data. You don’t need to do anything to make it happen, but you can prevent it.
Data collection
Through the modification of policy of privacy, the platform has communicated to European users that their personal data will be used for train models of intelligence artificial generative. The function is active automatically: anyone who does not intervene to deactivate it consents implicitly to treatment. A silent assent typically used by other digital platforms in similar situations.
The scope of the collection is very extensive if you consider that LinkedIn will draw from all of it historical available, starting from 2003year of birth of the professional social network. Personal and professional data will enter the training systems: name, profile photograph, experiences current and previous jobs, training path and skills declared, and then they will be collected post published or shared, articles written, i comments left under other people’s content, contributions to public discussions, responses to surveys, reporting problems. Everything that turns out publicwhile the only data excluded concern private messages, salaries and payment methods.
How to oppose
To prevent your information from being fed to artificial intelligence, it is necessary deactivate an active default setting. With a first method it is sufficient to log in to your profile in the app at Settings (gear symbol) or on the site at Settings and privacy (in the drop-down menu), choose Data privacy and then Data to improve generative AI: Here you can deselect the option Use my data to train AI models for content creation active by default.
An alternative is to compile a online form provided by LinkdIn itself. In the drop-down menu, under the item Select the type of objection or restriction requestthe option must be chosen Opposition to processing for training AI models for content creation.
The opposition can be given both by already registered users and by new ones, but with a clarification indicated by the Privacy Guarantor Italian which has published a dedicated information sheet. The objective is to facilitate the exercise of right of opposition foreseen by European regulation on the protection of personal data, and specifies that the opposition given after November 3 prevents the provision of data published after that date. It can only stop the collection of new updates, but do not cancel what has already been acquired.
The guarantor is however verifying the compliance of LinkedIn’s decision to regulation EU on privacy, checking the correctness of the opposition mechanisms, the type of personal data collected and the validity of the legitimate interest. Precisely through the latter, LinkdIn allows processing without explicit consentprovided that the user is offered the opportunity to object. It’s not the platform that has to ask, but the user who has to refuse. And the more time passes, the more material is provided to one machine and not to a person called to evaluate a CV.
