The UN Committee of Experts on International Cooperation in Tax Matters has supported consistent integration of the Subject-to-Tax Rule (STTR) into the 2025 UN Model Tax Convention Commentary and adopted a five-chapter guide outline on artificial intelligence in tax administration, covering responsible adoption, practical applications, risk management, and taxpayer safeguards.
At the fifth meeting of its 33rd session, the United Nations (UN) Committee of Experts on International Cooperation in Tax Matters supported the consistent inclusion of the Subject-to-Tax Rule (STTR), where appropriate, throughout the Commentary to the 2025 UN Model Tax Convention (E/C.18/2026/CRP.12).
The report outlines a six-part work plan to update the UN Model Tax Convention from 2026 to 2029, focusing on international tax issues affecting developing countries. It prioritises integrating the Subject-to-Tax Rule (STTR) into treaty provisions and commentaries to protect source-country taxing rights. It also proposes a broader, computation-based definition of a “low level of taxation” rather than relying on the narrower concept of “special tax regimes.”
UN Model Tax Convention update
The UN Tax Committee established the Subcommittee on the Update of the United Nations Model Tax Convention (co-coordinated by Wanjiru Kiarie and Antoine Guici) at its 31st session in October 2025. Its mandate focuses on addressing technical gaps and updating treaty provisions with a primary emphasis on issues most critical to developing countries.
Mandate, workplan & key timelines
The Subcommittee must recommend by April 2027 whether to prepare a 2029 update to the UN Model Tax Convention. Proposals for substantive revisions are due by autumn 2028, with final reviews and approvals scheduled for spring 2029.
Phase One workstreams
- Workstream 1 (STTR integration): Update treaty provisions and Commentary to reflect the Subject-to-Tax Rule (STTR).
- Workstream 2 (definitions): Review key treaty definitions and clarify terms that cause uncertainty in practice.
- Workstream 3 (source rules): Clarify income sourcing concepts such as “paid by,” “arising in,” and “deemed to arise,” particularly for passive income and cross-border services.
Phase Two workstreams
- Workstream 4 (profit attribution): Review rules for attributing profits to permanent establishments under Article 7 and its Commentary.
- Workstream 5 (PE concept & nexus): Consider updates to permanent establishment definitions to address non-physical presence and emerging digital business models.
- Workstream 6 (other issues): Address additional technical issues based on country demand and available resources.
Phase Two sequencing
Workstreams 4 and 5 will enter formal discussion after the 33rd session in October 2026, taking into account the pending work of the Digitalised and Globalisation Economy (DGE) and Transfer Pricing (TP) Subcommittees, particularly their evaluations of profit attribution.
Workstream 1: Deeper focus on the subject-to-tax rule (STTR)
The Subcommittee proposes amendments to the UN Model Tax Convention Commentary to align with Article 1(3) STTR, which protects source-state taxing rights when income is subject to low taxation in the residence state. The proposed changes include clarifying the STTR’s application across relevant treaty articles, introducing an optional provision based on a broader “low level of taxation” test without requiring connected parties, and amending Article 29 (Limitation of Benefits) to replace references to “special tax regimes” and remove connected-party restrictions.
Tax administration and artificial intelligence
The Subcommittee on Tax Administration and AI was mandated at the 31st session to draft a Practical Guide to Implementing Artificial Intelligence for Tax Administrations. The Committee adopted the detailed five-chapter outline presented at its 33rd session as outlined in document E/C.18/2026/CRP.16.
Detailed overview of the five chapters
Chapter 1: Introduction, purpose, scope, and context
Introduces the guide’s audience, objectives, and the role of AI in tax administration, particularly in developing countries. It distinguishes rules-based automation, machine learning, and generative AI, highlighting the need to integrate existing digital infrastructure, data, skills, legal frameworks, and governance systems.
Chapter 2: Key enablers for responsible AI adoption
Examines the essential requirements for adopting AI responsibly, including reliable technology, quality data, skilled staff, legal authority, and effective governance. It emphasises system interoperability, human oversight, workforce training, and appropriate procurement strategies.
Chapter 3: AI use cases by tax administrations
Explores AI applications in tax policy, taxpayer services, compliance, dispute resolution, and internal operations. Examples include identifying unregistered taxpayers, providing chatbot assistance, assessing audit risks, managing tax disputes, forecasting revenue, and estimating tax gaps. It also highlights the need to monitor taxpayers’ use of AI.
Chapter 4: Safeguards, risks, and mitigation strategies
Addresses the risks of AI use, including privacy breaches, data quality problems, bias, and errors. It emphasises data protection, human oversight, transparency, and taxpayers’ rights to explanations, challenges, and legal appeals, with safeguards proportionate to the significance of decisions.
Chapter 5: Conclusions and future developments
Highlights AI’s potential to improve tax administration efficiency while preserving human oversight and accountability. It encourages continuous testing, learning, and adaptation as AI technologies and their associated risks evolve.
