LinkedIn enters AI training

LinkedIn is building a marketplace for people to train artificial intelligence models, pushing the professional network into a fast-growing corner of gig work. (businessinsider.com) Business Insider reported on April 14 that LinkedIn has started listing early roles that pay as much as $150 an hour for tasks like testing chatbots and rating model answers. The report said the effort is aimed at fields including coding, finance, and other specialist work where companies want verified expertise. (businessinsider.com) This work sits inside the artificial intelligence training pipeline: companies hire humans to label data, compare outputs, and tell a model which answer is better. Data-labeling firm Snorkel describes the process as tagging raw text, images, audio, or video so a model can learn patterns and make predictions. (snorkel.ai) LinkedIn is entering a market that investors have already priced like core artificial intelligence infrastructure. Mercor raised $100 million at a $2 billion valuation in February 2025, while Surge AI was reported in July 2025 to be seeking up to $1 billion at a valuation above $15 billion and later discussed a round at at least $25 billion. (techcrunch.com) (finance.yahoo.com) (bloomberg.com) The pitch for LinkedIn is straightforward: it already has identity, resumes, work histories, and employer networks that startups have had to assemble from scratch. That gives it a built-in way to match a lawyer, accountant, or software engineer to a model-testing job that needs real domain knowledge. (businessinsider.com) The risks are also familiar to the sector. Axios reported in April 2024 that researchers found 550 gigabytes of exposed data, including artificial intelligence training material and employee credentials, on servers tied to a government contractor, underscoring how training pipelines can leak sensitive information. (axios.com) Large contractor pools have also brought labor scrutiny. Thomson Reuters said in January 2026 that data-enrichment work now spans collecting, curating, annotating, labeling, and evaluating data under a mix of employment models, while Scale AI said in March 2025 that it was being investigated by the U.S. Department of Labor over contractor pay practices. (thomsonreuters.com) (siliconangle.com) LinkedIn also arrives with its own recent privacy baggage around artificial intelligence. In September 2025, multiple outlets reported that LinkedIn expanded plans to use member data for generative artificial intelligence training and gave users an opt-out, adding another layer of sensitivity as it seeks to recruit people into the training economy itself. (pcmag.com) (pcworld.com) If LinkedIn can turn its professional profiles into a trusted labor pool, it will be selling more than job listings. It will be selling proof that the human side of artificial intelligence can be sourced, screened, and managed inside one of the internet’s largest work platforms. (businessinsider.com)