LinkedIn under fire for training AI models with user data behind the scenes

The service started to abuse users' data before making its actions public

Serving tech enthusiasts for over 25 years.
TechSpot means tech analysis and advice you can trust.

Privacy Violation 101: LinkedIn, the popular business and employment-focused social platform owned by Microsoft, appears to be positioning its users as data fodder for its generative AI algorithms. The company has initiated its AI training processes, but only notified users after the move was well under way.

LinkedIn recently confirmed that it is using user data to power its generative AI features. In a blog post by LinkedIn's SVP and General Counsel, Blake Lawit, the company outlined new user agreements and FAQs to inform users of these changes. However, it appears that LinkedIn began training its AI without first obtaining explicit user consent.

The Sunnyvale-based company's updated user agreement includes details about content recommendations, moderation practices, and the new generative AI features. The updated agreement will take effect in November.

In addition, LinkedIn has introduced a new privacy policy to clarify how user information is utilized in the development of its products and services, including AI-generated content. The policy specifies that LinkedIn collects, processes, and uses posts, articles, language preferences, and any feedback previously provided by users.

The company claims it is working to minimize the presence of personal information in data used to train AI. Through the use of "privacy enhancing" technologies, personal details can be redacted or removed from these datasets, we're told.

// Related Stories

• Game developer's demand that employees have naked sauna sessions sparks controversy
• Oracle's Larry Ellison advocates for AI-enhanced surveillance to promote "best behavior"

According to LinkedIn's new FAQs on AI, the platform leverages machine learning to suggest new article ideas. Users may provide personal data as input to a generative AI feature, which then uses these details to produce its AI-generated content. The generated draft can be edited or revised before being published, the company said.

It appears LinkedIn began utilizing users' data, including employment history and personal details, well before announcing these policy changes. However, users can now opt out of this unrequested feature by navigating to their account settings and selecting the new "Data for Generative AI Improvement" option. European users will be exempt from automatic data scraping for AI training "until further notice," according to the company.

In an attempt to justify this breach of privacy and trust, Lawit has framed LinkedIn's actions as part of a rapidly evolving technology and business landscape. However, many users, who rely on the network for professional connections and job opportunities, have not responded positively to the news. As one user remarked, "I used to think LinkedIn was less evil than other social networks. I guess I was very wrong."