Artificial Intelligence in Combating Tax Evasion: A Critical Comparative Analysis of the French and British Experiences and Its Applicability to the Algerian Tax Administration

Abstract

Tax evasion remains a significant loss of public revenue in both advanced and developing economies, and more and more tax administrations are using artificial intelligence (AI) in their control functions. This article explores two of the most institutionalized AI-powered fiscal enforcement models worldwide: the French Ciblage de la Fraude et Valorization des Requêtes (CFVR) program and the UK’s HM Revenue and Customs (HMRC) Connect system. The study presents a critical comparative analysis of their institutional origins, quantitative performance and legal-ethical constraints, drawing on documentary and secondary sources, before assessing the conditions under which comparable tools could be meaningfully adapted to the Algerian tax administration. The findings show that European systems are based on the progressive integration of several interoperable data sources, including fiscal, cadastral, banking, and (increasingly) social media data, and they produce measurable results (albeit methodologically limited) in terms of revenue recovery. A direct evaluation of Algeria’s digital tax architecture, through cross-checking with official sources from the Ministry of Finance and the DGI, finds that while the Jibayatic declaration and payment system has achieved a true national rollout, there is no legal requirement for electronic invoicing and no standardized tax-gap indicator by which the extent of evasion can be monitored, as is the case in France and the UK. The article concludes that a direct transposition of the CFVR and Connect models is neither realistic nor recommendable in the short term and instead promotes a phased, context-sensitive roadmap based on Algeria’s existing declarative infrastructure and the introduction of an e-invoicing mandate.