Browsing by Author "Alammari, Shada"

Now showing 1 - 1 of 1

Restricted
AI and Human Judgment in Financial Reporting: A Comparative Study of Accuracy, IFRS Compliance, and Professional Perceptions
(Saudi Digital Library, 2025) Alammari, Shada; Bosa, Iris
The accelerating integration of artificial intelligence (AI) into financial reporting processes has generated both promise and unease within the accounting profession. Contemporary developments in natural language processing (NLP) and generative AI, including large language models (LLMs), have made it feasible to produce complex financial outputs, such as statements of profit or loss (P&L) and statements of financial position (SFP), directly from structured or semi-structured data inputs. This technological shift raises fundamental questions about the quality, reliability, and compliance of AI-generated financial statements, particularly in contexts governed by International Financial Reporting Standards (IFRS) or national adaptations such as Saudi Arabia’s SOCPA-endorsed framework. While human accountants operate with judgement, professional scepticism, and regulatory awareness, AI operates through probabilistic pattern recognition. The central issue is whether such outputs can be relied upon for decision-useful financial reporting without undermining accountability and compliance. This dissertation addresses these concerns through a mixed-methods approach, combining (i) a simulation-based comparison of AI-generated and manually prepared financial statements derived from a reconstructed trial balance of Saudi Telecom Company (STC), and (ii) a survey of Saudi accounting professionals evaluating the perceived reliability, compliance, and adoption potential of AI in financial reporting. The simulation evaluates three key dimensions: numerical accuracy, IFRS compliance with respect to presentation and classification, and structural integrity, all benchmarked against a human-prepared standard. The survey extends this technical assessment by capturing professional perceptions of AI’s ability to complement or substitute for human judgement, as well as perceived risks such as regulatory non-compliance, loss of auditability, and the potential for novel misstatements.
28 0