Artificial intelligence (AI) is rapidly reshaping transfer
pricing (TP) practice and enforcement. Tax authorities increasingly use
AI-driven risk models for audit selection while multinational enterprises
(MNEs) apply machine learning (ML) to automate benchmarking, documentation, and
operational TP monitoring. This article reviews the emerging literature and
legal context to assess a key question: is AI in TP a risk or an opportunity?
It draws on OECD reporting, peer-reviewed studies, and practitioner and
regulatory materials to identify measurable opportunities including faster
benchmarking and documentation, predictive compliance controls, and more
consistent application of TP policies across jurisdictions. It also isolates
material risks such as opaque ‘black box’ outputs, data and algorithmic bias,
unsettled regulatory treatment under the EU AI Act and data protection law, and
expanding litigation and disclosure disputes. The article concludes that AI is
a net opportunity for TP governance only when deployed as decision support
under robust human-in-the-loop (HITL) safeguards with explainability, audit
trails, and periodic bias review sufficient to preserve the arm’s length
standard and procedural fairness.