New LinkedIn AI Data Policies Raise Concerns
LinkedIn, the Microsoft-owned professional networking platform, has recently updated its Terms of Service, sparking discussions about user data privacy and AI training practices.
In a blog post on Wednesday 18th September, General Counsel Blake Lawit outlines the changes, which clarify how LinkedIn handles user information, particularly in the context of training generative AI models.
LinkedIn has now stopped scraping UK user data, but is still scraping US user data by default. The UK's International Commissioner's Office revealed that LinkedIn pulled the plug on sourcing AI training data from the country's LinkedIn users after the regulator raised legal concerns.
However one significant aspect of this update is that user data from the US and several other regions will now be used by default to train LinkedIn’s AI models, without explicit user permission.
According to the company, these models may be enhanced not only by LinkedIn’s own technology, but also by other Microsoft services.
And users’ data could be utilised for various generative AI applications, such as providing writing suggestions.
For those concerned about their privacy, LinkedIn offers an opt-out option. Users can navigate to the “Data Privacy” section, select “Data for Generative AI Improvement,” and toggle off the “Use my data for training content creation AI models” setting.
Meta and X (formerly known as Twitter) are among platforms that, like LinkedIn, want to use content posted on their platforms to help develop their generative AI tools.
However, it’s important to note that opting out will not retroactively affect any training that has already occurred.
Interestingly, LinkedIn’s updated policy does not apply to users in the European Union or the UK due to stringent General Data Protection Regulation (GDPR) laws, which protect user data from being used without consent.
Despite LinkedIn’s assurances about transparency and user control, recent reports suggest that the company may have already begun using user data for AI training prior to announcing these changes.
A report from 404 Media claims that this practice was already in place, raising questions about the timing of the policy update.
These developments come on the heels of Meta’s recent admission that it has been training its AI on users’ data from Facebook and Instagram since 2007, also without explicit permission.
Unlike LinkedIn, Meta offers no opt-out option for most users, although Brazilian and European users can opt out due to legal rulings and GDPR protections.
As tech giants navigate the complex landscape of user data and AI, the tension between innovation and privacy continues to grow, leaving many users questioning the control they truly have over their own information.
