AIToday
Large Language ModelsAI Business & IndustryOpenAI BlogPublished: Jul 27, 2026, 22:00 JST

AI lets workers cross job lines, OpenAI study finds

AI lets workers cross job lines, OpenAI study finds

3 Key Points

  1. What happened

    OpenAI's new report, based on over 800,000 U.S. ChatGPT messages, finds that 16.8% of work-related messages and 43.5% of occupation-specific messages involve tasks tied to another occupation.

  2. Why it matters

    This 'task crossover' shows AI is expanding what workers do, letting a small-business owner draft copy or a marketer troubleshoot a website without waiting for specialists. The pattern is especially strong for customer experience (77% outside tasks), designers (75%), and HR (69%).

  3. What to watch

    The study suggests jobs may reorganize as tasks shift, but whether this leads to lasting change hinges on how firms adapt. Small businesses show more crossover—outside-task share drops from 18.9% for 2–5 seat workspaces to 16.3% for over 100 seats.

Not sure about something? Ask the AI

Summaries like this, in your inbox every morning.

Context & Analysis

OpenAI's analysis of ChatGPT usage reveals that AI is changing not just how tasks are done, but who does them. The numbers—16.8% of work-related messages and 43.5% of occupation-specific ones involving outside tasks—suggest a broad shift in role boundaries. The examples, like a small-business owner doing financial analysis or a marketer troubleshooting a site, illustrate this change in concrete terms. The study frames this as 'task crossover,' where workers borrow activities from other professions, potentially reducing the need for handoffs.

The report also shows that crossover is uneven across roles and firm sizes. Customer experience, design, and HR workers are heavy 'borrowers,' while engineering tasks are the most 'supplied' outward, appearing in 7.4% of other workers' messages. Marketing stands out in both directions, as marketers take on outside tasks (24.3%) and marketing tasks spread widely (8.9%). Smaller businesses see more crossover, likely because they lack specialist teams, giving workers an incentive to use AI as a generalist tool.

As a first installment in OpenAI's 'Work at the Frontier' series, this research will be followed by more data-driven insights. The stakes are significant: if task crossover becomes a lasting pattern, it could reshape job descriptions and career paths before official labor statistics catch up. However, whether this leads to broader organizational change depends on how companies choose to embrace or restrict these AI-enabled expansions of work.

FAQ
What is 'task crossover' in the study?
It's when AI users perform work traditionally associated with another occupation. For example, a marketer might troubleshoot a website, a task usually done by a developer, based on their ChatGPT usage.
Which occupations show the most crossover?
Customer experience workers (77%), designers (75%), and HR workers (69%) have the highest shares of occupation-specific messages involving outside tasks. Engineering supplies tasks that other fields adopt most.
How does business size affect crossover?
Smaller workspaces show more crossover: average users in 2–5 seat spaces have an 18.9% outside-task share, falling to 16.3% in over 100-seat spaces. This suggests AI helps where specialist resources are scarce.

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • Salesforce bets on Headless 360 beyond its own UISiliconANGLE AI · 1h ago
  • Google's Gemini hacked three firms, Google only confirms after WSJTechCrunch AI · 1h ago
  • Lossy self-improvement, not RSI: Anthropic sees no dramatic accelerationInterconnects (Nathan Lambert) · 4h ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleSingapore's Acrab unveils AI chip claiming 7.5× speed edge over Mac Mini M4 Pro