As artificial intelligence continues to reshapejobs and theworkplace as a whole, predicting its effects on the future of work has become more challenging than ever. Still, new research from OpenAI offers a fresh take on how these changes are unfolding. The July 2026 study revealed that ChatGPT users were regularly turning to the chatbot to complete tasks that fall outside the scope of their official occupations, a phenomenon the AI startup is calling “task crossover.”
What Is Task Crossover?
OpenAI defines “task crossover” as when workers use artificial intelligence to complete tasks that typically fall outside the bounds of their job and cross over into other roles. This does not include generic activities like writing emails or scheduling, but occupation-specific work like debugging code or creating marketing materials.
In the full report, the authors concede that their methodology contains some flaws. Even so, OpenAI’s research hints that AI is altering the traditional responsibilities of existing roles and potentially threatening several professions, serving as another reminder that tech workers need to upgrade their skill sets for the AI-driven workplace.
OpenAI’s Methodology
OpenAI analyzed more than 800,000 messages sent by ChatGPT Business users in the United States. They self-reported their occupations through their individual ChatGPT accounts, spanning eight fields:
- Customer Experience
- Design
- Engineering
- Finance
- Human Resources
- Legal
- Marketing
- Sales
To determine whether ChatGPT users were going beyond their typical responsibilities, OpenAI created a metric called “task crossover,” which refers to workers performing tasks traditionally done by another occupation. The company made sure to distinguish between specialized tasks and “generic” ones that are common across jobs, such as scheduling or writing. This way, OpenAI could zoom in on “occupation-specific messages” and identify when workers were using its chatbot to complete tasks beyond their roles.
What Did OpenAI’s Work Study Find?
More ChatGPT users are focusing on tasks outside the scope of their usual work, though it’s not the same across the board. Here’s what OpenAI discovered when it analyzed the data based on occupation and company size.
ChatGPT Users Are Performing Tasks Outside Their Occupations
OpenAI found that around 43.5 percent of occupation-specific messages involved tasks outside a user’s regular duties. Breaking it down by occupation, about 77 percent of these messages were from customer experience workers, 75 percent from designers, 69 percent from human resources workers, 56 percent from legal workers and 53 percent from marketers.
Certain professions are clearly using AI to perform duties usually handled by other positions, with nearly half of all ChatGPT messages reflecting this overreach. And not all occupations are impacted equally by this shift.
Marketing and Engineering Are Impacted the Most
Out of all the fields OpenAI looked at, engineering and marketing were the most exposed. At least 17 percent of messages in all occupations were related to engineering, and at least 18 percent were related to marketing. In other words, engineering and marketing were the fields that most frequently saw tasks performed by workers outside of those disciplines.
This trend may not be all that surprising, considering the range of coding assistants and low-code platforms available today. Not to mention that AI agents are able to solve increasingly more complicated software problems. Meanwhile, generative AI tools like ChatGPT have become adept at creating written content, images, soundtracks and videos — all the elements needed to produce physical and digital marketing materials.
Basic tasks in particular sectors are becoming much more accessible, thanks to AI. But some businesses are taking advantage of these capabilities more than others.
Smaller Businesses Are More Likely to Witness This Trend
The smaller the business, the more likely it is to experience task crossover. In this case, OpenAI defines small businesses as those with 2 to 5 seats, while large businesses consist of more than 100 seats. On average, about 18.9 percent of ChatGPT users completed tasks outside their roles in small businesses, dipping slightly to 16.3 percent at larger businesses.
It’s not a huge difference, but it’s enough to suggest that task crossover is a more common occurrence in smaller workplaces. After all, leaner organizations and startups may not have the same resources as bigger companies to support distinct, specialized teams. For workplaces with double- or even single-digit headcounts, employees are typically expected to be more flexible in picking up tasks, so naturally they may turn to AI to help manage their workload.
This pattern occurs in larger companies as well, raising questions about how all tech workers could be impacted moving forward.
What Does This Mean for Tech Workers?
OpenAI’s findings align with the general trend of more workers embracing AI. In fact, managers are beginning to base individual performance on AI usage, incentivizing employees to deploy the technology as much as possible. It stands to reason, then, that workers would use ChatGPT to handle a wider variety of tasks in hopes of ramping up their productivity.
But is this shift reaching the point where workers’ career prospects are at risk? If ChatGPT users are now completing simpler tasks typically done by other specialists, younger workers may have fewer entry-level opportunities and be forced to explore alternatives to thetraditional white-collar path. At the same time, previous studies on AI’s impact on jobs have revealed murky results, so it may be too much to say that AI is taking jobs right now.
An undeniable reality, though, is that employees who are literate in AI may have an advantage as workplaces become increasingly automated. A 2026 study by Anthropic reached a similar conclusion, suggesting that regular Claude users developed the skills employers coveted in the age of AI. Rather than AI itself eliminating jobs, those with the necessary skills would simply take over their colleagues’ tasks and roles, especially as managers begin to employ smaller teams who oversee networks of agents.
To avoid becoming expendable, tech workers may want to consider mastering crucial AI skills, whether it’s through employer upskilling programs, online courses or personal projects. There’s still no clear-cut relationship between artificial intelligence and mass job losses, but workers can take steps to adapt as artificial intelligence reshapes existing roles.
Why OpenAI’s Study Should Be Taken With a Grain of Salt
That said, it’s far too early to start panicking in light of this study’s limitations. The authors themselves note that the study analyzes messages sent by ChatGPT Business users, excluding ChatGPT Enterprise users entirely. It also covers just eight occupations, so OpenAI admits that it is “not representative of the entire U.S. workforce.”
Another major shortcoming is that the focus is on messages rather than workers’ actual progress. It suggests that workers are branching out to cover different types of tasks, but doesn’t say whether they’re actually becoming more productive or delivering higher-quality work as a result of using ChatGPT. The authors admit that they didn’t even determine whether a task could’ve been accomplished without AI in the first place.
In effect, it’s impossible to glean concrete insights from this study on whether ChatGPT use is impacting workers’ employment. OpenAI’s analysis could indeed serve as an early warning for tech workers who are unfamiliar with AI tools, but more research is needed to gauge whether AI is truly changing the nature of jobs and eliminating roles outright.
What is the main finding of OpenAI’s July 2026 study?
In a July 2026 study, OpenAI found that 43.5 percent of occupation-specific messages were related to tasks beyond that user’s typical job duties. In other words, nearly half of ChatGPT users rely on the chatbot to work on tasks traditionally handled by other specialists. Analyzing the data by profession, 77 percent of these messages were sent by customer experience personnel, 75 percent by designers and 69 percent by human resources workers.
Which fields are most frequently exposed to task crossover?
Engineering and marketing contain tasks most frequently performed by workers from other fields, according to OpenAI’s research. At least 17 percent of messages in every occupation studied were associated with engineering, and at least 18 percent were associated with marketing. This trend may reflect advances in tools like coding assistants and content generators, potentially signaling more drastic changes coming to these two disciplines.
What are the main limitations of OpenAI’s study?
The authors of the study note several flaws that weaken the credibility of their results:
- The study excludes messages from ChatGPT Enterprise users, resulting in a much smaller sample size.
- The study only analyzes eight occupations, failing to provide a comprehensive assessment of the broader American workforce.
- Messages are the focus of the study, rather than concrete metrics that measure productivity or performance.
- The authors didn’t determine beforehand whether tasks could be completed without AI, making the impact of ChatGPT use unclear.
Does OpenAI’s study prove that AI is causing job losses?
No, OpenAI’s study simply describes how some ChatGPT users collaborate with the chatbot, but it doesn’t deliver any definitive findings on whether AI is affecting workers’ employment or career prospects. It may highlight the importance of AI literacy and upskilling, but more research is needed to discern AI’s true role in the current job landscape.
