Machine Learning Security focuses on protecting the integrity, confidentiality, and availability of ML systems — from securing training pipelines against data poisoning to detecting adversarial examples that manipulate model outputs in production fraud, autonomous, and clinical AI applications. As organisations operationalise ML in regulated industries, the convergence of AI risk management frameworks (NIST AI RMF, EU AI Act) with traditional cybersecurity controls is creating a new discipline that Zero Trust architecture must extend to model registries, inference APIs, and data pipelines.
Related: AI & ML Security · CNAPP · Cloud Security Management · GDPR · Pharma