OpenAI says more workers are using ChatGPT to do other people's jobs

A recent analysis by OpenAI shows that more people are using ChatGPT to complete tasks outside their own professional domain. The company analyzed over 800,000 work-related ChatGPT messages and found that 43.5% of occupation-specific queries involved tasks from other professions. OpenAI calls this phenomenon 'task crossover.'

Specifically, marketing and engineering tasks are the most frequently cross-processed areas. Users leverage ChatGPT to handle work that previously required expert intervention, such as contract review, data analysis, and website troubleshooting. OpenAI believes these data are early signals that professional roles are shifting, even if job titles or descriptions have not yet caught up.

OpenAI notes that the task crossover effect is more pronounced in smaller companies. In small businesses, non-specialists are particularly inclined to use AI for marketing tasks. This is likely because small businesses lack dedicated expert teams, requiring employees to wear multiple hats, and AI fills that gap.

This finding suggests that AI tools may be helping small businesses acquire professional capabilities at lower costs, thereby altering traditional models based on specialized division of labor.

OpenAI used the U.S. occupational database O*NET to classify tasks, which maps activities to standard occupational profiles. To focus on occupation-specific tasks, OpenAI excluded common general tasks such as writing, summarizing, and scheduling.

It should be noted that this analysis is based on ChatGPT message data and may not fully represent all work scenarios. Additionally, task classification relies on O*NET's predefined framework, which may not capture the nuances of emerging or hybrid occupations.