The U.K.’s data protection authority has confirmed that LinkedIn, owned by Microsoft, has temporarily halted the processing of user data for training AI models. Stephen Almond, executive director of regulatory risk at the Information Commissioner’s Office (ICO), stated on Friday, “We are pleased that LinkedIn has taken our concerns seriously regarding its data practices for training generative AI models using information from U.K. users. We welcome LinkedIn’s announcement that it will pause this practice until further discussions with the ICO have taken place.”
Privacy advocates had early on noticed a subtle update LinkedIn made to its privacy policy following public outcry over its data usage for AI training. The company added the U.K. to the list of European regions where users do not have the option to opt out, asserting it is not processing data from local users for this purpose. LinkedIn’s general counsel, Blake Lawit, commented in a recent blog post that, “Currently, we are not using member data from the European Economic Area, Switzerland, or the United Kingdom for generative AI training, and we will not provide an opt-out option in these regions until further notice.”
Previously, LinkedIn outlined that it wasn’t processing data from users in the European Union, EEA, or Switzerland, where the General Data Protection Regulation (GDPR) is mandated. However, since U.K. data protection laws still align with the EU framework, privacy experts expressed dissatisfaction when it was revealed that LinkedIn did not extend similar treatment to U.K. users.
The Open Rights Group (ORG), a U.K.-based digital rights nonprofit, aired its disapproval of LinkedIn’s decision by filing a new complaint with the ICO concerning consentless data processing for AI. The group also criticized the regulator for not preventing further breaches concerning user data usage.
Recently, Meta, the parent company of Facebook and Instagram, resumed processing data from its U.K. users to train its AI models after lifting a prior pause. This means U.K. users with accounts must once more actively opt out if they wish to prevent Meta from using their personal data for its AI algorithms. Despite the ICO raising concerns about Meta’s data practices, the regulator has yet to intervene as the ad tech giant continues its data harvesting.
In a statement released on Wednesday, ORG's legal and policy officer, Mariano delli Santi, emphasized the danger of allowing powerful platforms to handle personal information without sufficient oversight. He asserted that platforms should be required to obtain explicit consent from users rather than relying on an inadequate opt-out system. He remarked, “The opt-out model is ineffective in safeguarding our rights; it is unreasonable to expect the public to monitor multiple companies that may misuse our data for AI training.” He added, “Opt-in consent isn’t just a legal necessity; it’s a common-sense requirement.”
We are currently awaiting responses from the ICO and Microsoft regarding these issues and will update this report accordingly.
Update: LinkedIn spokesperson Leonna Spilman stated in an email, “At this time, we are not enabling training for generative AI on member data from the European Economic Area (EEA), Switzerland, or the United Kingdom. We look forward to our ongoing constructive discussions with the ICO.” Regarding its approach to user data for AI training, LinkedIn further emphasized, “We believe our members should have control over their data, which is why we are providing an opt-out option for training AI models used in content generation in the applicable regions. Our services have always used some degree of automation, and we are transparent about how users’ data can be utilized. Today, many users seek assistance in crafting resumes, writing compelling LinkedIn profiles, or communicating with recruiters about new career opportunities. Ultimately, our generative AI services aim to provide users with a competitive advantage in their careers."
Ranking
