Artificial Intelligence Assets and Training Data under IFRS: Challenges and Application Prospects in Iraq

This research aims to address the accounting gap arising from the inadequacy of IFRS, particularly IAS 38, to accommodate the nature of artificial intelligence assets and training data, within the context of emerging economies with a specific application in Iraq. The study adopts a mixed-methods approach (quantitative and qualitative), including a survey of 178 accounting and auditing professionals, academics, and technical experts in Iraq, as well as a case study on the Bank of Baghdad. Results reveal a clear methodological gap in IAS 38 (mean perception 4.04/5), high practical challenges (4.29/5) – most notably the difficulty of distinguishing between research and development phases and allocating cloud costs. The proposed solutions (functional control, composite digital assets, hybrid measurement) received broad acceptance (4.31/5). The Bank of Baghdad case study confirms that applying the proposed model adds intangible assets worth IQD 3.67 billion and increases net income by approximately 3%, compared to the current treatment of writing off all costs. Main conclusions: a conceptual gap in current standards, an urgent need for an independent standard for digital assets or enhanced expanded disclosure as a transitional phase, and the current lack of readiness of the Iraq Stock Exchange. Recommendations: the Central Bank of Iraq and the Iraq Securities Commission should issue temporary local guidance, invest in specialized training programs, and conduct further large-sample studies.