ETHICS OF ARTIFICIAL INTELLIGENCE USE IN ACCOUNTING IN KAZAKHSTAN: SERVICES, ACCOUNTING-CYCLE STAGES AND A LEGAL RISK MATRIX

Abstract

The digitalisation of accounting in Kazakhstan is evolving from basic document automation to the use of artificial intelligence (AI) systems capable of recognising primary documents, classifying transactions, detecting anomalies and generating analytical explanations. Despite this transformation, professional responsibility remains with the accountant, while the legal framework is expanding through sectoral legislation, the Law “On Artificial Intelligence” and the Digital Code. The purpose of this article is to develop an applied ethical and legal model for the responsible use of AI in accounting in Kazakhstan, considering specific AI services, stages of the accounting cycle and regulatory requirements. The methodology combines a review of academic publications (2019–2026), analysis of Kazakhstan’s legislation, comparison with international principles of trustworthy AI, functional mapping of AI applications and an expertanalytical risk matrix developed by the author. Rather than measuring the actual frequency of violations, the matrix ranks AI-use scenarios according to four criteria: legal sensitivity, data confidentiality, algorithmic opacity and the impact on professional judgement. The study focuses on accounting processes where AI influences the reliability, confidentiality, explainability and accountability of accounting information. Six practical areas of AI application are identified: recognition of primary documents, transaction coding and classification, tax and accounting reconciliation, fraud monitoring, preparation of reporting explanations and audit support. The findings indicate that the greatest risks arise not from OCR-based document recognition but from automated tax interpretation, accounting classification and AI-generated explanatory notes to financial statements, where errors may compromise legal compliance and undermine professional accountability. The scientific novelty lies in integrating ethical principles with legal analysis and process-oriented accounting mapping within the Kazakhstani regulatory context. The practical contribution is a governance model based on data minimisation, mandatory audit trails, human validation of legally significant decisions, regular assessment of AIservice providers, and the formal allocation of responsibilities within organisational accounting policy.