Telfer Research Seminar Series - Devan Mescall

We provide evidence of critical factors supporting the successful adoption of artificial intelligence tools in organizations. Leveraging a unique cross-country dataset covering the investment and usage of AI tools by national tax administrations in 64 countries from 2016 to 2022, we use a stacked difference-in-differences framework to assess when and how the adoption of AI-based tools improved the key processes of tax administrative.  The setting provides a unique and robust opportunity to investigate the “jagged frontier” of capabilities provided by AI due to the alignment of goals and fundamental processes but variance of key organizational features such as governance, digital skills, resources and focus of AI tasks implemented across the sample of organizations. Our findings indicate that AI adoption is associated with significant increases in corporate tax revenue and improvements in administrative compliance, particularly in emerging economies, in countries with complex tax systems, and in those with digitally skilled populations. Cross-sectional tests reveal that benefits are strongest in countries with enforcement-oriented applications (fraud detection and risk assessment), while administrative and service-oriented uses show limited or no impact. By documenting how emerging technologies enhance regulatory enforcement capacity, our study contributes to our understanding of the importance of critical organization features in the success of AI investments as well as providing specific insights into the impact of AI on tax compliance, regulatory spillovers, and the role of digital monitoring in shaping taxpayer behavior.