OpenAI Research: Cross-Occupation AI Usage Doubles, Job Titles Remain Unchanged

On September 16, 2026, OpenAI Economic Research published follow-up research highlighting that how work is distributed and executed matters as much as AI tool accessibility. Analysis of over 1.5 million work-related ChatGPT messages from April to July 2026 confirmed that cross-occupation AI use beyond employees' existing job scopes has settled into a repetitive pattern rather than remaining a temporary experiment.

Cross-Occupation Usage Patterns and 'Borrowing Expertise'

When performing cross-occupation tasks, users generally used shorter prompts and more frequently provided examples or background information rather than requesting explanations or advice. They also often asked the AI to check or verify results. This suggests that users are leveraging AI to 'borrow expertise' from other fields by providing context such as documents or colleague information, rather than fully learning new domains.

Rising Repeat Usage Rates and Job Expansion

In a longitudinal study of approximately 6,200 workers, the share of cross-occupation tasks within job-specific AI activities nearly doubled, rising from 13.1% in April to 25.9% in July. The probability that workers who used cross-occupation tasks in the previous month would reuse them the following month was 23.6%, about three times higher than similar workers with no prior usage history (8.4%). This indicates that AI usage is becoming habitual.

Next-month reuse rates varied significantly by task type:

• Discussing products/services with customers: 54%
• Writing advertising or promotional copy: 44%
• Creating marketing materials: 37%
• Explaining financial information: ~15%

The overall average reuse rate for cross-occupation tasks was 18.5%. These differences are attributed to factors such as how naturally AI integrates into repetitive workflows, workplace norms, and perceptions of the severity of errors.

Implications: Changes in Work Content Precede Job Titles

The study emphasizes that the actual activity mix can expand even if job titles remain unchanged. A pathway is emerging where workers experiment with activities outside traditional roles, confirm AI's utility, and then incorporate these tasks into their regular work. Therefore, when developing AI adoption strategies, companies should consider work design that reduces friction, going beyond simply providing tool access.