Can artificial intelligence systems be valued as intangible assets?

Full Explanation

AI systems qualify as intangible assets when they meet the fundamental criteria: identifiability (the system can be separately identified, licensed, or sold), control (the entity controls access to the future economic benefits), and future economic benefits (the system generates or enhances revenue, reduces costs, or provides competitive advantage). Under IAS 38, internally developed AI systems can be capitalised once they pass the development phase criteria: technical feasibility is demonstrated, there is intent and ability to complete the asset, probable future economic benefits exist, adequate resources are available, and costs can be reliably measured. The research phase (exploratory data analysis, algorithm evaluation) must be expensed. The development phase (model training, testing, deployment preparation) can be capitalised. Valuation approaches for AI systems include the Cost Approach (reproduction or replacement cost of the model, data, and training), the Income Approach (present value of incremental cash flows attributable to the AI system), and the Market Approach (comparable transactions involving AI assets, though comparables are scarce). In M&A transactions under IFRS 3, acquirers must separately identify and value AI assets as part of purchase price allocation. The emerging ASU 2025-06 under US GAAP is expected to provide more specific guidance on AI asset capitalisation.

Related Glossary Terms

Related Questions