Artificial Intelligence (AI) has transformed modern cybersecurity by enabling intelligent threat detection, automated response, and predictive risk assessment. However, AI technologies are increasingly exploited by cybercriminals to launch adaptive, scalable, and autonomous cyber attacks. This research paper provides a comprehensive analysis of AI-powered cyber threats including automated malware generation, adversarial machine learning, AI-driven phishing, deepfake-enabled fraud, and prompt injection vulnerabilities. The study evaluates technical mechanisms, real-world incidents, organizational risks, economic implications, and national security concerns. Furthermore, advanced mitigation frameworks and AI governance strategies are proposed to enhance resilience against next-generation cyber warfare.