# What exactly did OpenAI find about how ChatGPT changes work?

According to OpenAI's research published on July 27, 2026, the company analyzed usage patterns of ChatGPT in professional settings and found a striking trend: users are consistently applying the tool to tasks that fall outside their official job descriptions. The study, which drew from anonymized usage data and user surveys, reported that over 60% of respondents said they used ChatGPT to complete work that would normally be handled by a different department or role. For example, a junior graphic designer reported using ChatGPT to draft client proposals — traditionally a task for account managers — while a customer support agent used it to generate SQL queries for data analysis.

This finding directly contradicts the dominant narrative that AI primarily automates or eliminates existing jobs. Instead, OpenAI's data suggests that AI is enabling workers to stretch their capabilities, taking on adjacent tasks that previously required specialized training or access to other teams. The research also noted that this 'role expansion' was most pronounced in small and medium-sized businesses, where employees often wear multiple hats already, but was also significant in larger enterprises with rigid job classifications.

# Why does this contradict the 'AI replaces jobs' narrative?

The dominant public discourse around AI, fueled by headlines from major outlets and think tanks, has focused on job displacement. However, OpenAI's findings align more closely with a smaller but growing body of economic research that views AI as a 'task expander' rather than a 'task eliminator.' Anthropic, in its own 2025 research on the economic impacts of AI, similarly found that AI tools were primarily used to augment human capabilities rather than replace them, with workers reporting higher productivity and broader skill usage.

The key distinction is that while AI can automate specific narrow tasks — like summarizing a document or generating boilerplate code — it simultaneously enables workers to take on higher-value, cross-functional work. According to OpenAI, the most common use cases for ChatGPT in the workplace were not simple automation but complex problem-solving, creative generation, and cross-domain translation. A software engineer might use ChatGPT to write user documentation, a marketer might use it to analyze customer data, and a finance analyst might use it to draft presentation narratives. This pattern suggests that AI is not shrinking the scope of human work but expanding it.

# Who benefits most from this role expansion?

OpenAI's research indicates that the benefits are not evenly distributed. The biggest winners are knowledge workers in roles that involve significant written communication, data analysis, and creative problem-solving — precisely the tasks that ChatGPT handles well. According to the study, workers in marketing, software development, data science, and customer support reported the highest rates of cross-role task expansion. These workers were able to take on tasks that previously required collaboration with other departments, reducing bottlenecks and speeding up project cycles.

Small business owners and freelancers also stand to gain disproportionately. The research showed that users in companies with fewer than 50 employees were twice as likely to report using ChatGPT for tasks outside their job description compared to those in companies with over 1,000 employees. This makes intuitive sense: smaller teams have fewer specialists, so the ability to 'borrow' expertise from an AI tool is more valuable. However, the research also noted a potential downside: workers in highly regulated industries like healthcare and finance were less able to take advantage of this expansion due to compliance and security concerns.

# What are the operational tradeoffs for organizations?

While the ability to expand worker capabilities sounds appealing, it introduces several operational challenges. First, performance evaluation becomes more complex. If a data analyst is now writing marketing copy, how do you measure their performance? Traditional job-specific KPIs become less relevant. According to OpenAI, early adopters are already experimenting with broader, project-based evaluation metrics rather than task-specific ones.

Second, training and upskilling needs change. If workers are expected to use AI to take on tasks outside their core expertise, they need guidance on how to evaluate the quality of AI output in unfamiliar domains. A software engineer who uses ChatGPT to draft legal documents needs basic legal literacy to review the output. According to the research, organizations that provided cross-functional training saw 40% higher satisfaction with AI tool adoption.

Third, there are accountability and liability questions. If a worker uses ChatGPT to generate a financial report that contains errors, who is responsible? The worker, the AI tool provider, or the organization? OpenAI's research did not address this directly, but it is a growing concern among legal and compliance teams.

# What should organizations do to prepare for this shift?

Based on OpenAI's findings, organizations should take three concrete steps. First, audit current job descriptions and identify which tasks are likely candidates for AI-enabled expansion. This is not about eliminating roles but about redesigning them to account for the new capabilities workers have. Second, invest in training programs that teach workers how to use AI tools effectively in cross-domain contexts. The research suggests that the best outcomes come from workers who understand both their core domain and the basics of adjacent domains.

Third, update performance management systems to reward outcomes rather than strict adherence to job descriptions. According to OpenAI, companies that have already adopted outcome-based evaluation report higher employee satisfaction and lower turnover. Finally, establish clear usage policies that address accountability, data privacy, and quality control. The research indicates that workers are more likely to use AI tools effectively when they have clear guidelines and support from management.

1. Prediction 1: By Q2 2027, at least three Fortune 500 companies will announce job redesign initiatives explicitly citing AI-driven role expansion as the catalyst.
2. Prediction 2: OpenAI will release a follow-up study within 12 months quantifying the productivity gains from role expansion, likely showing a 20-30% increase in output per worker for early adopters.
3. Prediction 3: The Society for Human Resource Management (SHRM) will publish new guidelines for AI-augmented job design by Q4 2026, acknowledging the end of rigid job descriptions.

• Insight 1: The 'AI replaces jobs' narrative is a distraction; the real story is how AI expands what workers can do, which creates new management challenges.
• Insight 2: Small businesses benefit most from AI-enabled role expansion because they lack specialist headcount, making AI a force multiplier for agility.
• Insight 3: Performance evaluation must shift from task-based metrics to outcome-based metrics as job boundaries blur.
• Insight 4: Regulated industries will lag in adoption unless AI tool providers build compliance-friendly features.
• Insight 5: The biggest losers in this shift are not workers but rigid organizational structures that cannot adapt to fluid roles.