OpenAI research finds 43% of ChatGPT work queries fall outside a user's own occupation

OpenAI published new research on July 27, 2026 showing that a large share of the work-related tasks people bring to ChatGPT don't match their job title — evidence, the company argues, that AI is starting to blur the boundaries between occupations rather than just automating tasks within them.

What's new

Analyzing occupation-specific ChatGPT messages, OpenAI found that 43.5% involve tasks outside the user's own occupation. The crossover is uneven across roles: customer experience workers pull in outside-occupation tasks 77% of the time, designers 75%, and HR professionals 69%. Marketing and engineering tasks turned out to be the most portable, showing up frequently in conversations from people who don't do those jobs. Smaller organizations show more crossover than larger enterprises, which OpenAI attributes to a narrower bench of specialists to lean on internally.

OpenAI frames the finding directly: "AI changes not just how work gets done, but who does what." The company argues this points toward task reorganization inside jobs rather than wholesale replacement of them, noting that "many jobs are likely to reorganize: these are jobs whose day-to-day tasks could change substantially." It also positions the tool as a stand-in for expertise that isn't otherwise available: "AI may be especially useful as a generalist tool where specialist resources are scarce."

Context

The research lands amid a broader industry and policy debate about AI's effect on employment, with OpenAI and other labs regularly publishing usage data to shape that conversation. Rather than the more commonly cited automation framing — AI replacing specific tasks within a role — this report leans into a labor-mobility framing: workers using AI to reach into adjacent skill sets they wouldn't otherwise have access to, particularly at smaller organizations that can't afford to hire a specialist for every function. That distinction matters for how the labor-market impact of generative AI gets modeled going forward.

Why it matters

If a large share of AI-assisted work is genuinely crossing occupational lines rather than just accelerating existing job tasks, it suggests generative AI's economic effect looks less like classic automation (fewer people needed to do the same job) and more like a redistribution of who can do what. That has direct implications for hiring, training, and org design — especially at smaller companies, where OpenAI's data shows the effect is strongest and where the ability to have one generalist employee credibly cover marketing, HR, or design tasks with AI assistance could reshape how lean a team can run. It's also a data point in the ongoing argument over whether AI mainly displaces jobs or reshapes them, with OpenAI, unsurprisingly, backing the reshaping story.