Here is a new paper written by the following set of authors: Manuel Hoffmann Harvard Business School, Sam Boysel Harvard Business School, Frank Nagle Harvard Business School, Sida Peng Microsoft Corporation, Kevin Xu GitHub, Inc. Here is the abstract:
Recent advances in artificial intelligence (AI) technology are showing great potential to complement labor-intensive tasks. While an emerging literature documents the broad productivity implications of AI, much less attention has been paid to how AI may change the nature of work itself. How do people, especially those in the knowledge economy, adapt the way they work when they start using AI? Using an open source software configuration, we study the effects of each level AI has on task allocation. We take advantage of environmental testing from the use of GitHub Copilot, a productive AI code completion tool for software developers. Using millions of jobs over a two-year period, we use a programmatic eligibility threshold to investigate the impact of AI technology on software engineer work allocation within an experimental regression design. We find that access to Copilot encourages such individuals to reassign work to their primary work of coding tasks and away from non-core project management tasks. We identify two basic mechanisms driving this change – the rise of independent rather than collaborative work, and the rise of exploratory rather than exploitative activities. The main effects are larger for people with relatively low ability. Overall, our estimates point to the great potential for AI to transform work processes and reduce organizational hierarchies in the knowledge economy.
By using the excellent Kevin Lewis.
Source link