Jeremy Bertomeu (Washington U St. Louis John M. Olin Business) et al. have posted “On Humans and AI: A Financial Reporting Dilemma” on SSRN. Here is the abstract:
Artificial intelligence presents unprecedented challenges in domains requiring nuanced ethical reasoning, particularly where objective training data is insufficient for complex moral judgments. This study investigates the ethical decision-making capabilities of large language models (LLMs) versus humans in an experiment featuring a board member’s dilemma over an untruthful, albeit well-understood by investors, reporting practice. Our research probes three fundamental questions: Do LLMs exhibit ethical judgments similar to humans, do they consistently apply ethical principles, and can they articulate the reasoning behind their moral choices? By comparing AI and human responses to an ethically ambiguous scenario, we aim to provide preliminary insights into the emerging role of artificial intelligence in navigating subjective decision-making processes. Our evidence contributes to our understanding of AI’s potential to augment or challenge human ethical judgment in high-stakes professional environments.