LinkedIn admitted Wednesday that it has been training its own AI using users’ data without explicitly seeking consent. Users cannot opt out of training that has already occurred. LinkedIn limits opt-out options to future AI training only.
In a blog post detailing updates effective November 20, LinkedIn general counsel Blake Lawit confirmed that amendments to the user agreement and privacy policy will specify how personal data fuels AI on the platform. The updated privacy policy now informs users that their personal data may be used to develop and train AI models, provide and personalize services, and gain insights using AI, automated systems, and inferences.
An FAQ explained that personal data could be collected during any interaction with generative AI features, including composing posts, changing preferences, giving feedback, or general usage. The data is stored until users delete AI-generated content. LinkedIn recommends using its data access tool for deleting or requesting to delete past activities’ data.
The FAQ warned that users providing personal data to generative AI features could see their data surfaced as output. Although LinkedIn claims it seeks to minimize such risks by using privacy-enhancing technologies, the platform restricts opt-out options to future data collection only.
To opt out of future AI training on LinkedIn, users can navigate to the “Data privacy” section in their account settings and turn off the option allowing collection of “data for generative AI improvement”. However, this does not apply to users in the European Economic Area or Switzerland, as stricter privacy laws already require explicit consent.
Users can also object to the use of their personal data for non-content-generating AI models via the LinkedIn Data Processing Objection Form.
LinkedIn, a company that previously committed to AI transparency, faces growing scrutiny as users grapple with the implications of their data being used to train AI. The platform’s updated user agreement warns about the potential inaccuracies and risks associated with AI-generated content, advising users to cautiously evaluate all information shared on LinkedIn.
For further information, visit the original article on Ars Technica.