TAX EVASION AND FINANCIAL FRAUD IN THE CURRENT DIGITAL CONTEXT

Preventing and combating phenomenon of tax evasion is a present concern of national governments due to the magnitude this phenomenon represents and because of the increasingly sophisticated techniques used by the authors in carrying out tax frauds. Evolution of tax evasion phenomenon at international level has acquired a profound technological character due to the increasingly elaborate methods. Illegal behaviour has some specific features that could be recognized easily by artificial intelligence models. They use real data in order to derive characteristics that could be identified in due time so that tax avoidant behaviour be identified and prevented. The use of forecasting models like logistic regression, random forests or decision trees in order to model tax avoidant behaviour shows having a good predictive power. Also, the use of the neural networks allowed scientists to calculate probability of an individual taxpayer that would attempt to evade taxes or commit other types of financial frauds. Scientific literature shows an increasing interest in using neural networks to detect and predict fraudulent behaviour in the fields of tax avoidance and financial domain. Cybercrime, cryptocurrency and blockchain were created in order to facilitate payments and help owner in accumulating wealth. Current landscape of financial frauds shows a different picture. Intracommunity frauds are more and more diversified. European Union and International bodies act together to prevent and combat fraud. Could these new technologies possess a real threat to the financial security of our transactions or encourage fraudulent behaviour? This paper tries to find the answer to this question