A recent study from Harvard Business School has offered new insights into how generative artificial intelligence (AI) could help managers—and particularly those in technical roles—rediscover the aspects of their work that they are most passionate about. By analyzing the activities of more than 187,000 software developers over two years, the researchers found that access to AI tools like GitHub Copilot didn’t just make developers more efficient—it shifted the nature of their work, giving them more space to focus on what they enjoy and excel at.
Becoming a manager typically means trading in hands-on work for more administrative duties, which not all professionals find fulfilling. This study, however, suggests that AI can help change that equation. “You get into a job because you love the core work,” said Frank Nagle, Assistant Professor at Harvard Business School and lead author of the study. “As you become more senior, you start doing more management work. Some people like that, but some people don’t. This is showing that AI helps people get that balance back closer to what they would prefer it to be.”
The research, titled Generative AI and the Nature of Work, was conducted by Nagle along with Manuel Hoffmann and Sam Boysel of the Laboratory for Innovation Science at Harvard. The team collaborated with GitHub software engineer Kevin Xu and Sida Peng, a senior principal economist at Microsoft, which owns GitHub. They focused their study on open-source developers on GitHub and used a natural experiment involving GitHub Copilot—a generative AI code assistant—to compare activity levels between developers who had access to Copilot and those who didn’t.
The data set spanned from July 2022 to July 2024 and revealed striking differences in behavior. Developers with access to Copilot spent 12% more time on core coding tasks and 25% less on project management and administrative work. Interestingly, the number of collaborators these developers worked with dropped significantly—from an average of 22 down to just 5. This points to a clear shift toward more independent, self-contained work styles, driven by AI assistance.
AI access also sparked more innovation. Developers began experimenting with new programming languages—usage went up by nearly 22%—and engaged in 15 new open-source projects on average. “If this is a tool that allows people to explore more, then that’s probably a good thing because we’re getting new ideas and new projects,” said Nagle.
These changes could have financial benefits too. The study estimates that increased engagement with programming languages could raise a developer’s earning potential by roughly $1,700 annually. When applied to the nearly 300,000 active open-source maintainers on GitHub, this could amount to an aggregate value of $468 million a year.
The impact was most pronounced among less-experienced developers. Those new to the field increased their time spent coding by 11%, more than double the 4.6% increase seen among more experienced peers. They also cut down project management time by 27%, compared to 14% for seasoned professionals. This indicates that generative AI may serve as a highly customized learning and productivity tool, helping newer workers build their core skills more quickly.
“It’s kind of like a Choose Your Own Adventure book,” Nagle said. “Everyone can choose the best path for them and their skillsets.”
The findings of this study underscore a broader trend: generative AI is not just transforming productivity—it’s redefining job satisfaction and opening new possibilities for how work is structured in knowledge-intensive industries. For managers and developers alike, AI could be the key to falling in love with their work all over again.
