OpenAI's Own Data Shows ChatGPT Is Blurring Job Boundaries at Work

Customer experience staff are writing marketing copy. Marketers are doing legal research. HR teams are debugging spreadsheets meant for finance. According to a new OpenAI report, this kind of role-crossing has become a defining pattern of how people actually use ChatGPT on the job.

The report, titled "Work at the Frontier: How AI is Expanding What People Do at Work," is the first in a new research series OpenAI is calling "Work at the Frontier." It draws on an analysis of more than 800,000 work-related ChatGPT messages sent by users in the United States, with tasks classified against O*NET, the U.S. government's standard occupational database.

The topline figure: 43.5% of occupation-specific messages involve tasks that belong to a different profession than the sender's own. OpenAI calls this "task crossover," and it excluded generic activities like writing, summarizing, or scheduling from the count, meaning the overlap is concentrated in more specialized, job-specific work.

Zoomed out across all work-related messages, not just occupation-specific ones, the crossover rate is 16.8%. But the pattern is uneven across professions. Customer experience workers top the list at 77%, followed by designers at 75%, HR professionals at 69%, legal workers at 56%, and marketers at 53%.

Marketing and engineering emerged as the two occupations whose tasks most frequently show up in other people's ChatGPT conversations, suggesting these fields have become de facto shared toolkits that non-specialists reach for regularly.

Small businesses feel it most

OpenAI found the crossover effect is more pronounced at small businesses, where dedicated specialist functions such as in-house legal counsel, brand designers, or HR departments are often absent. Employees there appear to lean on ChatGPT to cover gaps that would otherwise require hiring outside expertise or a specialist colleague.

The data suggests AI is not just automating tasks within a job, it's letting individual workers take on responsibilities that used to require a different job title.

Alongside the report, OpenAI introduced what it calls an "AI Jobs Transition Framework," positioning the research as a tool for understanding how job design and hiring might evolve as generalist AI use expands. The company frames the findings less as evidence of job displacement and more as a signal that work itself is being restructured around who can access which capabilities, not just who holds which title.

For employers, the implications touch hiring, training, and org design directly. If a single employee can competently draft legal language, run customer support scripts, and produce marketing assets with AI assistance, the traditional logic of narrowly scoped roles and separate specialist teams may need rethinking, particularly for smaller organizations without the headcount to staff every function separately.