Since ChatGPT’s debut less than three years ago, generative AI has started reshaping how people work, enabling many tasks to be completed faster or even automatically.
Researchers are now examining how generative AI might impact middle management roles.
Impact of Generative AI on Task Distribution Within Organizations
A team led by Harvard Business School professor Manuel Hoffmann studied this effect by analyzing how people interact with GitHub Copilot, a generative AI tool for coders.
From 2022 to 2024, the researchers observed 50,032 global software developers, half of whom used GitHub Copilot, and tracked over 2.4 million actions.
These actions were categorized as either core work (coding) or managerial work (project management).
The study found that Copilot users focused more on coding and less on managerial tasks—coding activity rose by 5%, while project management activity dropped by 10%.
Hoffmann noted: “Studying developers and how they used GitHub provided us with really detailed information about their work.”
Two main reasons explained this shift: First, gen AI helped developers work more independently, which reduced their need to collaborate or coordinate.
Second, faster coding allowed more time for exploratory tasks like learning or trying new methods.
These findings suggest that companies might flatten their hierarchies—not by removing all middle managers, but by reducing their numbers.
Managers who remain could spend less time on coordination and more time on meaningful work, including tasks typically done by contributors.
Changes in Middle Management Roles
As gen AI improves at scheduling and quality control, managers may have more room for innovation and direct involvement.
However, companies must address two major questions: what tasks should be automated, and who gains the most from automation?
Hoffmann recommends cataloging tasks—like his team did—to identify which duties can shift from humans to AI.
The study revealed that Copilot users sought less help from managers and peers and often worked in smaller teams.
Hoffmann explained: “Gen AI can, in some instances, replace a manager, mentor, or any other individual that junior employees may ask for help.”
Less skilled developers benefited more from Copilot, as they relied more on AI guidance and less on managerial support.
This reduces the burden on managers to train lower performers and increases productivity across teams.
It also allows managers to assess employee capabilities more effectively and guide them in using AI for improvement.
Although the research focused on coders, observing how all employees interact with gen AI can reveal broader areas for flattening hierarchies and increasing efficiency.
Hoffmann concluded: “We will likely see an increase in agility in companies that use gen AI. This will lead to the flattening of corporate hierarchies, which will help streamline productivity and reduce the need for so many middle management tasks.”
