LinkedIn may eye labeling

A leaked internal LinkedIn dashboard suggests the company is building an artificial intelligence data-labeling operation, according to screenshots posted by investor Turner Novak on April 12. (x.com) Data labeling is the work of tagging text, images, audio, or video so an artificial intelligence model can learn from examples or be graded after it responds. Companies buy that work when they need humans to rank answers, mark objects, check safety, or create specialized training sets. (appen.com) LinkedIn has not publicly announced a labeling product, and the leak does not show launch timing, pricing, or customers. LinkedIn did not immediately provide a public statement in sources reviewed for this article. (x.com) The idea fits LinkedIn’s existing business. Microsoft said in its fiscal 2025 annual report that LinkedIn makes money from Talent Solutions, Sales Solutions, Premium Subscriptions, and Marketing Solutions, and Microsoft said LinkedIn revenue rose 10% in fiscal first quarter 2025. (microsoft.com 1) (microsoft.com 2) Microsoft and LinkedIn have also been pushing deeper into artificial intelligence products for recruiters and professionals. In January 2025, Microsoft said LinkedIn had passed $2 billion in Premium subscription revenue over the prior 12 months, after adding more artificial intelligence features to paid products. (techcrunch.com) (microsoft.com) If LinkedIn moves into labeling, it would join a field that already includes specialist vendors and platform operators. Appen says it has almost three decades in data sourcing, annotation, and model evaluation, while Uber says its artificial intelligence services unit offers annotation and labeling across text, audio, images, and video. (appen.com) (uber.com) Uber’s move showed that a company with an existing marketplace and global operations stack could repurpose that machinery for annotation work. Uber said in November 2024 that its Scaled Solutions division would hire contractors in the United States, Canada, and India for both internal and external labeling projects. (techinasia.com) (uber.com) The competitive pressure has already been building. Appen reported fiscal 2025 revenue of $230.8 million, with growth tied to new project wins and generative artificial intelligence work, while private competitors such as Scale AI and Labelbox have pitched labeling plus software as a combined service. (quartr.com) (crunchbase.com) (cbinsights.com) For buyers, the attraction of a LinkedIn-backed operation would be less about basic tagging than about control over workforce identity, audit trails, and integration with hiring software. Those are the same enterprise concerns that already sit inside LinkedIn’s recruiting and subscription businesses. (microsoft.com) (techcrunch.com) The leak is still only a leak, but it points to one clear direction: large platforms are treating human feedback work as infrastructure, not just outsourced labor. Whether LinkedIn ships the product or not, that is the part competitors will have to price against. (x.com) (appen.com)