LinkedIn pays people to train AI

LinkedIn is testing a marketplace that pays professionals to train artificial intelligence systems, with some listings offering as much as $150 an hour. (businessinsider.com) The early roles focus on specialized work, not generic data entry: coding, nursing, finance, healthcare, and language tasks where a model’s answer has to be checked by someone who knows the field. LinkedIn confirmed to Business Insider that the product is in early testing. (businessinsider.com) In India, that top rate has been widely translated to about Rs 14,000 an hour, and local coverage says the jobs involve reviewing chatbot answers, labeling data, and judging whether code or domain-specific responses are accurate. (indiatoday.in) LinkedIn’s pitch is that this is “flexible, skill-based income” for members whose expertise can be turned into training data. The company has described the work as helping companies build “high-quality, human-labeled data for AI training.” (socialmediatoday.com) That work is the part of artificial intelligence that still needs people: humans ask questions, rate answers, spot mistakes, and supply examples so models learn what a good response looks like. Companies across the sector use that process to improve chatbots’ accuracy and safety. (builtin.com) The pay range also shows how much more valuable expert judgment is than basic annotation. Other AI training platforms now advertise generalist work from about $20 an hour, while specialist roles in coding, medicine, finance, and other fields can run far higher. (dataannotation.tech) (micro1.ai) LinkedIn is not entering an empty market. Business Insider reported that the company’s new marketplace would compete with specialist intermediaries such as Mercor and Surge AI, which already connect companies with freelancers who evaluate model outputs. (businessinsider.com) The timing fits LinkedIn’s broader argument that artificial intelligence is creating new categories of work even as it automates older tasks. LinkedIn data cited by the World Economic Forum in January said artificial intelligence had added 1.3 million jobs globally over the previous two years. (weforum.org) The catch is that these jobs are still in a pilot phase, with eligibility checks and screening before workers are matched to projects. For now, LinkedIn is treating human expertise itself as the product — and paying a premium when that expertise is hard to fake. (economictimes.indiatimes.com)