LinkedIn hit with lawsuit for training AI on private messages

You pay a monthly fee for a professional networking tool expecting career leads and extra privacy, not to have your private direct messages turned into textbook material for someone else's AI model.

That scenario is now at the center of a proposed class-action lawsuit filed in the U.S. District Court for the Northern District of California against Microsoft-owned LinkedIn. The lawsuit accuses the platform of unlawfully handing paying LinkedIn Premium customers' private messages over to third parties to train artificial intelligence models, alleging the company stealthily altered its privacy policies to hide what it was doing.

The complaint specifically zeros in on InMail messages, demanding $1,000 in damages per plaintiff on behalf of Premium subscribers.

Opted in by default, disclosed a month later

The dispute hinges on a timeline that plaintiffs argue is proof of foul play.

Back in August 2024, LinkedIn rolled out a new privacy setting that let members opt out of having their personal data fed into AI training pipelines. The catch was that the toggle was turned on by default for everyone, putting the burden entirely on users to discover the setting and switch it off.

Then, a month later, LinkedIn quietly updated its formal privacy policy. The revision stated that user data could indeed be used to train AI models and potentially shared with third-party partners. The lawsuit claims this retroactive rewrite was a deliberate attempt to cover the company's tracks after already violating contract terms and data privacy protections.

LinkedIn, for its part, is pushing back on the claims. A company spokesperson stated that LinkedIn is not using member messages to train models as alleged in the complaint.

Scraping user data for machine learning is an industry-wide flashpoint right now. Tech giants including Google, Microsoft, and OpenAI have all faced court battles over taking user data without direct consent, but LinkedIn's case hits a particularly raw nerve because it involves paying users and private communications.

Why it matters

For anyone holding a Premium subscription, this lawsuit is a blunt reality check. Paying for a platform doesn't automatically protect your private business communications from being treated as free training fuel. If you have an active account, you should head straight into your account settings and verify whether the AI training toggle is switched off.

For the AI industry, this case marks a dangerous escalation in the hunt for specialized training data. Public web scrapers have hit diminishing returns, making private business correspondence and networking chats look like gold to model builders trying to master professional communication. But private messaging is built on strict trust, and treating it as fair game risks breaking that foundation entirely.

We think LinkedIn's approach here looks remarkably clumsy. Relying on default opt-ins and following up with retroactive policy tweaks is exactly the kind of deceptive design that invites regulatory hammers and alienates paying customers. Even if LinkedIn successfully proves in court that private InMail contents were never ingested, the mere appearance of quietly siphoning user data has already dealt serious reputational damage.