Microsoft Research has released a paper that ranks 80 occupations by how closely their daily tasks align with the capabilities of large-language-model chatbots such as Bing Copilot. The study analysed 200,000 anonymised user exchanges with Copilot last year and calculated an “AI applicability” score for each role based on the proportion of tasks the system performed to users’ satisfaction. Interpreters and translators topped the list with a score of 0.49, followed by historians, passenger attendants, service-sector sales representatives and writers. Customer-service representatives, CNC tool programmers and telephone operators also feature among the 40 roles judged most exposed to generative AI. In contrast, hands-on and equipment-intensive jobs—including dredge operators, roofers, nursing assistants and phlebotomists—showed near-zero overlap with the technology. The research underscores an emerging divide between knowledge work and manual labour but stops short of predicting mass redundancies. “Our data show where AI can support tasks, not that it can fully perform an entire occupation,” senior researcher Kiran Tomlinson said, adding that the technology is more likely to reshape work than erase it outright. The findings arrive as companies across finance, technology and retail accelerate adoption of generative AI, a trend that consultancy Challenger, Gray & Christmas estimates was linked to more than 10,000 US layoffs in July. Microsoft’s study suggests that the most immediate pressure will fall on entry-level and information-processing roles, even as it argues that new positions may emerge alongside the software.