Understanding Adversarial Attacks Against Machine Learning & AI. As AI and Machine Learning become embedded in critical systems, they are creating new attack surfaces and security risks. Cyberweek Africa examines Adversarial Machine Learning (AML) attacks specifically designed to exploit vulnerabilities in ML models across their development, training and deployment lifecycle. Such attacks can manipulate model behavior, compromise sensitive information and potentially affect interconnected systems. The key message is clear: AI security must be built in from the beginning. Developers, cybersecurity professionals, decision-makers and risk owners must understand AI-specific threats and strengthen security throughout the ML lifecycle.