IOTA Digital Workshop on Using AI Solutions to Detect Tax Evasion
On 3–4 December 2025, IOTA successfully conducted its Digital Workshop on Using AI Solutions to Detect Tax Evasion. Organised under the IOTA 2025 Work Programme Activities, the two-day virtual event brought together over 250 tax administration officials, data specialists, and technical experts from across the IOTA membership to examine emerging applications of artificial intelligence in the detection and mitigation of tax evasion. The event was broadcast live via MS Teams.
The Workshop provided a platform to present concrete country initiatives, explore methodological and operational challenges, and exchange best practices for integrating AI-driven tools into tax administration processes. Participants discussed both strategic considerations and practical implementation issues, including data governance, model development, and cross-institutional cooperation.
On both days of the event, the Plenary Sessions focused on country-specific approaches to Initiatives and Projects on Using AI to Detect Tax Evasion. The first day featured expert presentations from IOTA members of Denmark, France, and Hungary.
- Denmark: “Use of AI against Taxcrime - The AML Prioritisation model” presented by Andreas Refshøj Brahe;
- Hungary: “Machine learning approaches for predicting NACE activity codes of Hungarian business entities” presented by Gergő Bence Mamuzsics;
- France: “France's Use of Artificial Intelligence in Detecting Tax Evasion” presented by Thibault Guillemat
Speakers presented national projects employing machine learning, predictive analytics, and automated risk detection. Their presentations highlighted technical aspects of model development, approaches to identifying behavioural anomalies, and early lessons learned from operational deployment.
The second day of the Workshop included contributions from Bosnia and Herzegovina and Spain.
- Spain: “Incomes level. How AI closes the information gaps” presented by Rafael Padilla;
- Bosnia and Herzegovina: “Detecting VAT Fraud Using AI” presented by Mario Blažević
These case studies demonstrated how AI-powered systems can enhance compliance monitoring and support targeted interventions. Presenters discussed challenges such as data integration, system scalability, and organisational readiness, while also outlining opportunities for further innovation and cross-country cooperation. Each day, participants actively exchanged ideas and experiences in smaller virtual breakout sessions and engaged in Q&A sessions.
All Workshop materials are available on the event’s webpage, and the video recordings will be accessible shortly on the same platform. The Workshop was facilitated by IOTA International Taxation Experts Massimo Morarelli and Alexandros Roukalis, whose guidance ensured a structured and effective exchange of knowledge throughout the event.
