LinkedIn Faces Lawsuit Over AI Training Data Privacy

Microsoft is in hot water this week after its business-focused social media platform LinkedIn was targeted with a proposed class action lawsuit.

The lawsuit claims that LinkedIn violated its contractual obligations by giving third parties access to the private messages of its Premium customers in order to build generative artificial intelligence (“AI”) models. “These communications contain extremely sensitive and potentially life-altering information about employment, intellectual property, compensation, and other personal matters, given its role as a professional social media network.”

The lawsuit claimed that “there was swift and harsh public backlash when it was publicly revealed that LinkedIn had unilaterally disclosed its users’ data for these purposes.” In response, LinkedIn quietly changed one of its privacy policies that same day to accommodate data sharing linked to AI. It also said that users could “opt out” of further disclosures for these reasons.

Plaintiff Alessandro De La Torre filed the complaint on Tuesday night in the Northern District of California on behalf of millions of LinkedIn Premium subscribers.

Premium customers claim that Microsoft’s LinkedIn, alleging that the business-focused social media network revealed they shared their private chats with third parties without their permission in order to develop generative AI models.

Last August, LinkedIn launched a privacy setting that allowed users to enable or prohibit the sharing of their personal information. 

LinkedIn agreed in September 2024 to cease its use of UK user data to train generative artificial intelligence (AI) models and in a “frequently asked questions” linked, opting out “does not affect training that has already occurred.” in response to concerns raised by the UK Information Commissioner’s Office.

At the time, the ICO stated it was ‘pleased’ with Redmond’s answer.

It was claimed that a choice was provided to opt out of the plan, however users are opted in by default, which means that the data will be used unless users are aware of it and choose to opt out.

In addition to the United Kingdom, LinkedIn had exempted clients in Canada, the EU, EEA, Switzerland, Hong Kong, and mainland China from having their LinkedIn data used “to train content-generating AI models.”

According to a proposed class action filed on Tuesday night on behalf of millions of LinkedIn Premium customers, the company secretly implemented a privacy setting in August that allowed users to enable or disable the sharing of their personal data.

This attempt to “cover its tracks” implies LinkedIn was fully aware that it had violated customers’ privacy and its promise to use personal data exclusively to promote and grow its platform in order to avoid public scrutiny and legal ramifications, according to the complaint.

The action was filed in San Jose, California, federal court on behalf of LinkedIn Premium customers who sent or received InMail communications and had their private information revealed to third parties for AI training prior to September 18.

It demands unspecified damages for breach of contract and violations of California’s unfair competition statute, as well as $1,000 per person for violating the federal Stored Communications Act.

LinkedIn stated, “These are false claims with no merit.”

A lawyer for the plaintiffs had no immediate reaction. The case was filed a few hours after US President Donald Trump launched a $500 billion joint venture between Microsoft-backed OpenAI, Oracle, and SoftBank to construct AI infrastructure in the United States.