Mert Demirer

A new study by assistant professor Mert Demirer and co-authors tracked software developers' work before and after they adopted AI tools. They found an explosive impact at the top of this funnel — coders created or edited almost 300% more files — but that boost was halved to 150% by the time they got to the number of discrete pieces of work submitted for review, and that in turn shrunk fivefold to a roughly 30% uplift in the number of full software releases.

Research, by assistant professor Mert Demirer, associate professor John Horton, Peyman Shahidi (PhD candidate), and co-authors models production as a sequence of interdependent steps and shows that it is this interdependence that determines the true extent of the gains enabled by AI. "We are seeking to understand the effect of AI at the overall system level, not as a one-off productivity tool applied on a task-by-task basis," said Shahidi.

"I will expect some impact on the legal profession's labor market, but not major," said assistant professor Mert Demirer. "AI is going to be very useful in terms of information discovery and summary," he said, but for complex legal tasks, "the law's low risk tolerance, plus the current capabilities of AI, are going to make that case less automatable at this point."

Experts say the transition to an A.I.-powered workplace is likely to be more gradual, in many cases occurring as new companies, built to exploit A.I., take market share from more established companies that are slower to embrace it. "Widespread adoption is going to happen at the new firms," said assistant professor Mert Demirer. "It's always the case that the smaller the production process, the more the process is easier to change."