More than half of organizations are using AI, by one measure, and business leaders are eager to leverage the technology to maximize efficiency.
Image: Shutterstock
Becoming a manager usually means spending more time on process and paperwork and less time doing what you love. Now, a novel study shows that generative artificial intelligence (AI) could give managers some balance back.
Researchers from Harvard Business School analyzed the activities of more than 187,000 software developers over two years to see how using AI tools changed their workdays. The study offers an intriguing finding: AI didn’t just help developers code more efficiently, it empowered them to approach their jobs differently and follow their interests, says HBS Assistant Professor Frank Nagle.
“You get into a job because you love the core work. And then, as you become more senior, you start doing more management work,” Nagle explains. “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.”
More than half of organizations are using AI, by one measure, and business leaders are eager to leverage the technology to maximize efficiency. The study is one of the first to show how AI can help reframe parts of individual jobs, particularly in management, the authors say. While the study focused on software development, they highlight generative AI’s potential to transform how work is divided and prioritized across other knowledge-intensive professions, suggests Nagle.
Nagle conducted the study, “Generative AI and the Nature of Work,” with Manuel Hoffmann and Sam Boysel, both postdoctoral fellows at the Laboratory for Innovation Science at Harvard. The scholars collaborated with Kevin Xu, a software engineer at the software collaboration platform GitHub, and Sida Peng, a senior principal economist at Microsoft, which owns GitHub.
Massive dataset of developer activities
Nagle and his team based their study on open source developer activity from GitHub, which allow them to analyze the impact of GitHub’s Copilot AI tool. Some open source core developers, called “maintainers,” were given free access to Copilot if the projects they worked on were above a ranking threshold, allowing for a comparison of those developers who were just above the threshold with those who were just below (and therefore did not get free access).
Open source software source code is produced by teams and distributed for free, a valuable resource that underpins many other technologies. Maintainers shoulder heavy administrative and managerial loads to orchestrate the myriad contributions from the growing community.
The team observed the developers weekly activity from July 2022 to July 2024. Their main finding: Developers with access to Copilot increased “core” coding activities by 12 percent over the non-Copilot group. They decreased their project management and administrative work by 25 percent.
Also read: How to understand AI’s potential impact on knowledge jobs
Less collaboration, more experimentation
Though developers are known to be highly collaborative, the study showed those with AI access engaged with others far less. They worked with an average of five collaborators in public projects, down 79 percent from the control group’s 22 collaborators.
With AI access, these developers “began working on smaller projects with fewer people involved and tasks that required less interaction,” Nagle explains.
Using Copilot also allowed individual developers more space for experimentation. On average, those with AI access increased their use of new programming languages by almost 22 percent while engaging with 15 new open source projects.
This finding suggests that AI can serve as a catalyst for innovation. “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,” Nagle says.
One exciting area to explore, Nagle says, is how these shifts apply at the team level. The findings suggest that workers in the future might pursue greater specialization, “where the people who want to write code, write code, and the people who prefer more administrative tasks can do more of that work,” he explains.
A Choose Your Own Adventure tool
More immediately, the study shows that specialization could have financial implications. The study determined that increased exposure to programming languages could increase developers’ earnings potential by about $1,700 per person or $468 million annually for the nearly 300,000 open source maintainers active on GitHub.
Newcomers to the field were poised to benefit most, as the researchers found AI had the most significant impact on relatively inexperienced developers. This group increased their time spent on coding by as much as 11 percent, compared with 4.6 percent for more established developers.
Similarly, the less-experienced developers reduced project management tasks by as much as 27 percent—doubling the 14 percent reduction for the more seasoned developers.
While some of these differences reflect common sense—for instance, more experienced developers may already feel comfortable with the balance of managerial tasks—it bodes well for generative AI more broadly as a “customized” learning and development tool. As Nagle says: “If you’re good in one thing, it makes it easier to be good in another thing.”
“You could certainly watch a YouTube video to learn,” he continues. “But would it be tailored to you? That’s one of the powerful things about this technology. It’s kind of like a Choose Your Own Adventure book. Everyone can choose the best path for them and their skillsets.”
This article was provided with permission from Harvard Business School Working Knowledge.
