AI trustworthiness under agency theory: a theoretical approach

Abstract

This study explores the expanding relationship between humans and AI, focusing on the role of AI trustworthiness in addressing Agency Theory issues. The research underscores the necessity of aligning AI autonomy with the potential risks and consequences of its decisions, noting the detrimental impact of bias and other negative outcomes. The core trustworthy AI attributes examined are Fairness, Accountability, Transparency, and Ethics. Five propositions are posited: 1) Trustworthiness facilitates human reliance on AI decisions. 2) Transparency reduces information asymmetry between humans and AI. 3) Fairness curtails AI bias and moral hazards. 4) Accountability enhances AI alignment with human objectives and 5) Ethical AI mitigates agency problems across all stakeholders. This research bridges AI trustworthiness with Agency Theory within Human-AI collaboration, offering theoretical insights. It highlights AI’s agency, discussing risk causes, effects, and mitigation strategies for managers utilizing AI as a tool, team member, or manager