OECD highlights data protection, transparency and skills challenges for use of Artificial Intelligence
On 18th September 2025, the OECD issued its report “Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Functions” exploring both the potential and the risks of AI adoption in tax administration. While AI is already widely used by tax authorities to enhance efficiency, compliance, and service delivery, the report underscores critical challenges related to privacy, fairness, and accountability.
Tax administrations have long utilised rule-based AI to detect tax evasion, improve services, and streamline decision-making. More advanced tools are now being used to analyse unstructured data such as social media posts and handwritten documents, as well as satellite imagery — for example, to identify undeclared property or assets. AI is also being used to pre-populate tax returns, answer taxpayer inquiries through virtual assistants, and assign risk scores based on transactional data, enabling tax agencies to better focus their resources.
See how AI VAT advice works already in VATCalc’s single appellation for calculations and returns.
Risks and challenges in the implementation of AI
The report discusses issues around privacy and the general public being willing to accept AI processes:
-
Data Protection and Privacy: Tax administrations handle vast amounts of sensitive data, making robust governance essential. Poor-quality or incomplete data can lead to inaccurate risk assessments and biased outcomes.
-
Transparency and Explainability: Complex AI systems, particularly machine learning models, can operate as “black boxes,” making it difficult for taxpayers to understand or challenge decisions. This threatens the legal rights of taxpayers and undermines the rule of law.
-
Trust and Voluntary Compliance: As tax systems rely on taxpayer goodwill, any perceived unfairness or opacity in AI use can erode trust, discouraging compliance.
See how Artificial Intelligence can work for VAT.
Hurdles to implementing AI
The report also notes operational challenges, such as:
-
Skills shortages in AI development and implementation.
-
High costs of adoption and infrastructure investment.
-
Outdated or inflexible legal frameworks that are not tailored to the unique needs of tax systems.
To move forward, the OECD calls for a systemic and integrated approach to AI in tax administration. This includes embedding AI into broader digital ecosystems, updating regulations (e.g., using “Rules as Code”), and ensuring transparency and human oversight. The OECD Forum on Tax Administration is piloting a framework based on principles for trustworthy AI, with the aim of helping tax authorities manage these challenges effectively. Findings from this initiative are expected in 2026.
