Cyber-security in the era of AI is a double-edged sword: machine-learning models now generate convincing spear-phishing, deepfakes, and self-mutating malware, yet the same techniques power real-time anomaly detection and autonomous incident response. But attackers also exploit vulnerabilities in models through prompt injection, data poisoning, and model inversion.
In this session we will talk about how to mitigate these risks with e.g. robust model governance and secure ML pipelines and how AI-driven defense mechanisms, including behavioral analytics and automated threat detection, can outpace traditional security approaches.
