AI-driven data breaches have emerged as one of the biggest challenges for corporations. New threats emerge daily in the form of personalized spear phishing messages, deceptive emails, deepfake voice or video, and disinformation. Research in this domain has so far focused primarily on what measures should be taken after a data breach; however, it remains unclear what capabilities are required to defend a company. To address the unknown and unknowable threats, the study presents how firms can develop AI-powered dynamic capabilities to tackle the next level of AI-generated threats.