OpenAI Uses 800,000 Messages to Expose Professional Boundaries: 43.5% of Work Inquiries Have Crossed Their Own Desks

OpenAI has brought the idea that 'AI is breaking down the walls between jobs' to quantifiable data for the first time. It examined over 800,000 ChatGPT messages related to work and found a startling statistic: 43.5% of job-specific queries actually point to tasks from another profession. The company calls this phenomenon "task cross-pollination"—people are no longer limited to what's on their title, but are using AI to step into the territory of other roles.

Tasks in marketing and engineering are the most frequently overlapping, like two threads that are often intertwined. Users are now handling tasks that used to be handed off to experts: reviewing contracts, analyzing data, and troubleshooting website issues—work that was once tied to legal, analyst, or operations roles—is now being handled by one person with a chatbot.

This trend is especially strong in small companies. They often can't afford a dedicated team of experts, so non-professionals are more willing to use AI to fill roles like marketing—without a dedicated team to back them up, AI becomes an easily available substitute. OpenAI believes these usage patterns are an early signal before changes in job titles and descriptions occur, indicating that reorganization of roles is already happening, even if the organizational charts haven't caught up yet.