Artificial intelligence (AI) security solutions are fundamentally built to distinguish between “safe” and “malicious” user activity by comparing user behaviour across environments to behaviour in similar environments. Unsupervised learning is the term used to describe the process whereby the system develops patterns without human supervision. “Deep learning” is a crucial tool for recognising fraudulent behaviour on some AI platforms, including Vectra. Deep learning is based on vast, linked networks of artificial neurons, which are inspired by the biological composition and operation of neurons in the brain. These neurons are arranged into layers, and each layer is connected to every other layer by a set of weights that change in response to incoming inputs.