Exclusive: Workers are crossing job boundaries with AI, OpenAI research shows

• If the trend persists, it could reshape hiring and job design — long before those changes show up in official labor market statistics

What they're saying: "The boundaries between jobs are likely already becoming more flexible due to AI," OpenAI chief economist Ronnie Chatterji tells Axios.

By the numbers: Roughly 44% of workers' occupation-specific ChatGPT requests involved tasks typically associated with another profession, after stripping out generic tasks like writing emails and scheduling meetings.

• The research draws on a sample of more than 800,000 work-related messages from U.S. business users.

Zoom in: Workers turned to ChatGPT to create marketing materials, troubleshoot software, perform financial calculations and explain regulations — tasks traditionally handled by specialists.

• Customer service workers, designers and HR professionals showed the highest rates of such crossover work, with roughly three-quarters of their occupation-specific prompts involving tasks more commonly associated with another profession.

Yes, but: The authors caution that their analysis does not determine if AI is creating new cross-occupation work versus helping workers perform responsibilities they already had.

• Nor does it measure whether those requests translate into higher productivity, comparable work quality or different hiring decisions.

What to watch: Among typical ChatGPT users, nearly 19% of work-related requests at the smallest businesses crossed occupational boundaries, versus roughly 16% at larger firms.

• That's consistent with workers at smaller firms using AI to fill expertise gaps when specialists aren't readily available, the report says.

The bottom line: AI may change the work people do before it changes the number of people who do it.