Source: Ciconte, Rozario & Urcan (2026) — “Using AI to Identify Exogenous Shocks and Conduct Archival Accounting Research”
This is one of those risks that sounds abstract until you imagine explaining it after the fact. Then it suddenly becomes very concrete, very expensive, and very difficult to hide behind a slide deck.
There is a particular kind of panic that only research can create. Then you realise you need a clean research design, a valid shock, the right dataset, a defensible method, and ideally a small miracle. So of course the idea of using AI as a research assistant is tempting. Maybe ask it to also reassure you that doing a doctorate was a good life choice 😭 This paper tested that temptation in a very direct way. The researchers used AI to identify regulatory shocks in securities rules, test whether those shocks affected voluntary disclosure, and then draft academic papers from the findings. AI was reasonably useful at identifying possible shocks and giving the researchers a starting point. The draft paper was not ready for a top accounting journal, which is a polite academic way of saying “please do not submit this unless you enjoy rejection.” But it saved time. It gave the researchers something to interrogate. That is the good version of AI in research. More like a fast assistant who gives you a messy first map, and then you still have to check whether the roads exist.
In plain English, that is why the result matters beyond the chart. It changes where people should look, what they should question, and which comfortable assumption probably needs to be retired.
My takeaway: the danger is rarely the dramatic thing in the headline. It is the quiet gap between knowing a risk exists and assigning someone to do something about it. Very unglamorous. Very important.