Will AI Take Your Job? (cacm.acm.org)

CEOs from Fiverr to Amazon have publicly warned that AI will displace jobs, and emerging research shows those warnings aren’t purely rhetorical: studies using text-mining of the International Standard Classification of Occupations (ISCO) and real-world usage data point to concentrated exposure in knowledge work. Indiana University researcher Ali Zarifhonarvar finds “professionals” (95 occupations) and “technicians and associate professionals” (60 occupations) face the greatest “full impact” risk—meaning a majority of core tasks could be automated—especially in media (writers, journalists, advertising), legal support, finance analysts, and even software and data roles. Microsoft’s analysis of 200,000 Copilot interactions flags high exposure for interpreters, sales reps, writers, customer service agents, and reporters, while freelance-platform studies show demand for writing and software gigs falling roughly 30% and 20% respectively. The technical takeaway: AI now handles data processing, text/code generation, analysis, and advisory tasks fast and cheaply, changing which job tasks are automatable rather than only eliminating routine physical work. Prognoses range from the World Economic Forum’s estimate of 92 million jobs lost by 2030 to Anthropic’s Dario Amodei suggesting up to 50% of white-collar roles could vanish in years. Jobs requiring physical presence, manual dexterity, or nuanced human judgment—crafts, on-site healthcare, construction, plant operators—remain most sheltered for now. That gap implies substantial reskilling/upskilling needs and a likely rise in inequality unless policy and training scale quickly.