From Human Judgment to Machine Learning: A Philosophical Study of Knowledge Creation in Modern Accounting

Abstract

Background: The integration of artificial intelligence (AI) and machine learning into accounting and auditing has transformed the creation, validation, and application of professional knowledge, raising questions about the legitimacy of algorithmic knowledge in accounting practice.

Objective: This study examines the epistemological shift from human judgment to algorithmic knowledge in the implementation of Financial Accounting Standards (Standar Akuntansi Keuangan [SAK]) and Quality Management Systems (Sistem Manajemen Mutu [SMM]).

Methods: This study employs a qualitative-interpretive approach through conceptual analysis and an examination of practitioners’ perspectives on the reliability, transparency, explainability, and trustworthiness of AI-assisted decision-making.

Results: The results of the study show that AI can improve efficiency, predictive accuracy, and analytical capabilities in modern accounting practices but, at the same time, poses epistemological challenges in the form of limited algorithmic transparency, limited explainability, and the potential erosion of auditors’ professional skepticism. Although the conceptual analysis indicates that AI can positively affect efficiency and predictive accuracy, it also raises epistemic challenges related to transparency, accountability, and the preservation of professional skepticism. The shift from judgment-based knowledge to data-driven inferential knowledge has transformed the traditional epistemic authority held by accountants and auditors. The application of SAK and SMM in the accounting profession has implications for defining professional responsibilities in maintaining the quality of financial reporting and audit processes.

Conclusion: This study proposes a human–machine epistemic partnership, in which AI functions as an analytical augmentation tool while accountants and auditors remain the primary epistemic authorities.