How AI Is Expanding What People Do at Work

OpenAI analysis of 800,000+ U.S. ChatGPT work-related messages finds that 43.5% of occupation-specific AI use involves tasks traditionally belonging to a different job category — a pattern the researchers term 'task crossover.' Customer experience workers (77%), designers (75%), and HR workers (69%) show the highest rates of borrowing tasks from other occupations. Marketing and engineering tasks travel farthest across organizational boundaries, with marketing work appearing in 8.9% of messages from workers outside marketing. Task crossover is more pronounced in smaller organizations (18.9% outside-occupation share in 2–5 seat workspaces vs. 16.3% in 100+ seat workspaces), suggesting AI functions as a generalist resource where specialist functions are scarce. OpenAI frames these usage patterns as a leading indicator of occupational restructuring that traditional labor-market statistics will capture only later.