Economists are increasingly embracing AI tools to streamline tasks like data cleaning and grant writing, with some reporting a fivefold increase in time for analytical work. At Dartmouth College, Paul Novosad noted he could now focus more on research questions thanks to AI automation, while ETH Zürich’s Elliott Ash admitted the productivity boost made him “want to work more.” However, AI’s role in producing groundbreaking research remains limited. David Yanagizawa-Drott of the University of Zürich described outputs from a project generating 1,000 economic studies as “decent” but not Nobel-worthy, and journals like the *American Economic Review* report only 25% of submissions disclose AI use, with no clear quality shift.

Beyond productivity, AI is expanding the scope of economic inquiry. Researchers can now analyze previously intractable datasets, such as corporate earnings calls or zoning regulation impacts, thanks to tools that simplify qualitative data processing. The Bank for International Settlements highlighted AI’s forecasting power, citing a model that predicted inflation trends in 2021–2022. These advancements suggest AI is democratizing complex analyses but struggles with depth.

Peer review is also evolving. The AI tool Refine, tested by top journals, detects errors in one-third of submissions, even those reviewed by human editors. Creator Ben Golub noted its potential to “plug holes” in porous review processes. Yet risks persist: academics have submitted poorly written AI-generated referee reports, raising concerns about over-reliance. Economists must balance AI’s efficiency with vigilance to avoid complacency.

The discipline faces a pivotal question: Can AI’s error-detection capabilities outpace human negligence? While tools like Refine show promise, the long-term impact hinges on whether researchers prioritize critical thinking over convenience.