Linkedin is said to have shared private messages for AI training
A class action lawsuit against Microsoft’s subsidiary LinkedIn is currently causing a stir in the USA. Premium users (i.e. paying users with subscriptions) accuse the platform of passing on private messages to third-party companies without explicit permission in order to train AI models.
As the BBC reports, LinkedIn is said to have “quietly” introduced a new privacy setting in August 2024 that automatically included user accounts for data sharing for AI training (“opt-out”). A month later, according to Reuters, a discreet adjustment to the privacy policy was made, which for the first time mentioned the use of data for AI training purposes.
What is particularly controversial is that the FAQ later added that a subsequent opt-out has no influence on data usage that has already taken place. The plaintiffs see this as an attempt to “cover their tracks,” as the court documents show.
LinkedIn counters: “False claims without foundation”
The lawsuit was filed in California, on behalf of LinkedIn Premium users. It seeks $1,000 in damages per affected user for alleged violations of the US Stored Communications Act, as well as other unspecified amounts for breach of contract.
LinkedIn, which has over a billion users worldwide – almost a quarter of them in the US – strongly denies the allegations. A company spokesperson told the BBC: “These are false claims without basis.”
The geographical component is interesting: As an email from Linkedin to its users shows, data sharing for AI purposes has not been activated in Great Britain, the European Economic Area, and Switzerland. Users in these countries are therefore probably not affected by the case.
There is a lot at stake for LinkedIn: The company generated revenue of $1.7 billion from premium subscriptions alone in 2024. The platform says it is experiencing rapid growth in premium subscribers, which is also related to the increasing integration of AI features.
