Title: Adoption of Artificial Intelligence in Accounting and Management Earnings Forecast

Authors: Congcong Li (Duquesne University), Mark (Shuai) Ma (University of Pittsburgh), Michael Shen (National University of Singapore), Yucheng (John) Yang (The Chinese University of Hong Kong)

Abstract: Using a novel dataset of online job postings, we identify firms that adopt artificial intelligence in accounting based on the hiring of accountants with artificial intelligence skills. Our stacked difference-in-differences regressions find that firms’ management earnings forecast accuracy significantly increases after they adopt artificial intelligence in accounting. Cross-sectional analyses suggest that artificial intelligence in accounting improves forecast accuracy by helping firms more timely incorporate information into their forecasts and mitigating errors in management forecasts due to economic uncertainty. But, artificial intelligence in accounting does not seem to mitigate intentional biases caused by management’s economic incentives. Further, financial analysts appear to understand the implications of artificial intelligence in accounting for management earnings forecasts. In addition, we provide a number of tests to rule out several alternative explanations. Overall, our study helps us better understand the consequences of adopting artificial intelligence in accounting to firms’ voluntary disclosure behavior.