AI가 연구 논문을 죽일 것인가?

Imagine taking a macroeconomics paper and adding a little button at the end “Press this button to update this paper with the latest macro data.” All of a sudden you have multiple papers rather than one, and no single canonical version. It is the latter versions, not created directly by the authors, that people will look at. Imagine adding another button, to either micro or macro papers “Please rerun these results using what the AI thinks might be five other different yet still plausible specifications.” Then you have more papers yet. Ultimately, why not just build a “meta-paper,” using AI, to answer any possible question about the subject area under consideration. This meta-paper would allow the reader, using AI, to make many sorts of modifications and additions to the basic work. The meta-paper also would allow the reader to add new data, to run additional robustness checks, and to do whatever else you might think of. Once again, the canonical version of the paper evolves away. A researcher might spend a significant part of his or her career building such a meta-paper. Imagine a meta-paper, or sometimes I call it a “box,” devoted to answering questions about say fiscal policy, minimum wage hikes, or maybe the Industrial Revolution. Fed researchers would spend their entire careers, not writing papers, but improving the Fed’s “box” that answers questions about monetary policy and also prudential supervision. Who will be good at doing such things? Is it the people today who become the top economists, or not? Will it be a highly decentralized endeavor, or, given the compute and team work requirements, a highly centralized one? Economics is going to change a lot, as will many of the other sciences. It is funny, and tragic, how much some of you are still obsessed with writing and publishing “papers.”