This paper investigates whether the use of AI by securities regulators in countries within the European Economic Area enhances their oversight efficiency and, in turn, the financial reporting quality of firms under their supervision. We identify regulators’ AI usage through public procurement contracts. Using a difference-in-differences design, we find that AI usage improves oversight efficiency, as reflected in increases in the number of firms subject to enforcement actions, both in absolute terms and relative to the number of completed examinations. The improvement in oversight efficiency strengthens with greater AI usage. Interestingly, the effect is similar for AI used for financial reporting oversight and for other AI applications. Regulators’ number of employees and total employee costs do not decrease, suggesting that AI usage by regulators complements, rather than displaces, human labor. Consistent with strengthened monitoring, firms exhibit a lower likelihood of accounting fraud following their regulators’ adoption of AI. The deterrence effect is stronger for small firms and those with low market-to-book ratios, which were less likely to be targeted for financial reporting investigations under the traditional, non-AI regime.

zlynne@ceibs.edu