Examining Algorithm Aversion in Jurors’ Assessments of Auditor Negligence: Audit Data Analytic Exception Follow-Up with Artificial Intelligence

New technologies offer benefits to auditors. However, resource constraints may prevent follow-up testing on all of the large volume of exceptions typically identified by audit data analytics (ADA) population tests, potentially increasing liability if an undetected misstatement was identified as an exception. Holding the use of ADA constant, we examine if how auditors follow up on exceptions (artificial intelligence or human auditors select a sample for further testing) affects auditor liability. We also examine two potential interventions to mitigate liability when auditors use ADA: a hypothetical ADA standard and experienced auditor pre-review and approval of the selected sample of exceptions for follow-up testing. Results from two experiments suggest higher auditor liability when AI selects the sample for follow-up testing, an effect mediated by perceived auditor causation. Of the interventions we examine, only experienced auditor pre-review mitigates liability when using ADA, an effect mediated by perceived misstatement foreseeability