If AI Helps You Do Someone Else’s Job, Should Your Role Change?

OpenAI analyzed more than 800,000 work-related ChatGPT messages. Axios reports that, after generic requests were removed, about 44% of occupation-specific prompts involved work usually tied to another profession. At the smallest businesses, the crossover share was nearly 19%, versus about 16% at larger firms. For a five-person company, that can be genuinely useful. The office manager can get oriented on a payroll formula. The designer can troubleshoot a software issue. Work does not sit for three days waiting for a specialist the company does not have on staff. But “ask ChatGPT” can also become a cheap way to widen a job. The employee gets the first draft, the uncertainty, and the blame if a polished answer is wrong. The analysis does not tell us whether the output was good, whether productivity improved, or whether people were already doing this work. A manager should be able to say which new tasks are now part of the role, what training comes with them, and where the answer still goes to a paid expert. If none of those change, this may not be time saved. It may be specialist work pushed onto whoever already has the fullest plate. Would you change someone’s training, pay, or scope