AI Security Platforms protect machine learning models, generative AI applications, and their data pipelines from adversarial attacks, misuse, and unauthorized exposure. Core capabilities include model red-teaming and adversarial testing, runtime guardrails against prompt injection and jailbreaks, output filtering for sensitive data leakage, shadow-AI discovery across the enterprise, model-provenance and supply-chain validation, and policy enforcement against approved-use boundaries. The category emerged as organizations deployed LLM-powered applications and realized that traditional AppSec and network controls miss model-specific threats; leading vendors include Protect AI, Robust Intelligence, Lakera, HiddenLayer, Cranium, and Prompt Security.