Artificial Intelligence on Both Sides of the Barricades: How Generative Technology and Behavioral Analytics are Rewriting the Rules of Cybersecurity

The article is devoted to the analysis of current trends in the field of cybersecurity of financial organizations in the context of active development of generative technologies. The evolution of cyberfraud methods is studied where artificial intelligence is used both by attackers and by protection systems. The scientific novelty of the article lies in proposing an applied model for setting anti-fraud thresholds that simultaneously takes into account the expected prevented damage, the cost of excessive blocking of the legitimate operations, and the expenses for additional verification. Unlike descriptive threat analysis, the proposed approach allows for the translation of cyber risks into monetary metrics and their use to make management decisions between blocking, additional verification, and seamless operation passing. The work considers current threats including deepfakes, generative phishing, and malware, and analyzes current methods of countermeasures based on behavioral analytics and real-time systems. Special attention is paid to regulatory changes and technological solutions including the Unified Biometric System and SIS (state information system) Antifrod. The paradoxes of modern cybersecurity have been analyzed: despite the high efficiency of anti-fraud systems (99.7% of blockages), the damage caused by fraud continues to grow. The author proposes a comprehensive protection strategy that combines neural network scoring, risk quantification, and institutional mechanisms.