LinkedIn was reportedly caught training models with user data before updating its policies
LinkedIn opted in users in the US to data collection for AI model by default and without updating its privacy policy, prompting privacy concerns. The company also failed to correctly specify whether UK-based users were subject to the data collection process, spurring further confusion.
Several LinkedIn users in the US recently noticed a new toggle on the platform's settings, followed by a disclaimer that LinkedIn is using personal data and users' content to train generative AI models. As expected, the toggle is on by default, meaning that users must opt out if they don't want LinkedIn and its affiliates (Microsoft, for instance) harvesting their personal data and content for model training purposes. Also unsurprising was the fact that the move was done without any update to LinkedIn's privacy policy.
Independent journalism website 404 Media was the first outlet to pick up the news. The website obtained a statement from LinkedIn confirming that it hadn't updated its policy and would do so shortly. LinkedIn's privacy policy has since been updated; although not without further controversy.
In a blog post announcing the updates to its privacy policy update, LinkedIn clarified that users from the European Union (EU), the European Economic Area (EEA), and Switzerland would not see the toggle in their settings because the company is not scraping data from users in these regions. This spurred the nonprofit Open Rights Group (ORG) to call on the ICO to take action against companies scraping users' data without opt-in consent.
ORG also highlighted that the UK's Information Commissioner’s Office (ICO) has not yet responded a complaint it filed about Meta's data processing, where much like noyb– European Center for Digital Rights, ORG argues that Meta has violated the GDPR because it has no legitimate interest to override the interests of users whose data is being harvested. Additionally, Meta provides no opt-in consent mechanism, giving users a right to object, rather than respecting their right to consent to having their information harvested. The company recently announced it would resume training on data from users in the UK after working with ICO to simplify its opt-out mechanism.
ORG's call to action seemed especially relevant since the original blog post said nothing about the UK being exempt from the data harvesting process. Regardless, UK-based users did not even have the option to opt out of the process because the toggle was also unavailable to them. LinkedIn has quietly edited the blog post to include the UK among the regions exempt from data scraping.
To opt out of having personal data scraped, users have to navigate to the "Data Privacy" section of their settings menu, click on Data for Generative AI improvement, and turn the Use my data for training content creation AI models toggle off. Users wanting to object to data processing not covered by the toggle must fill out this form.
