Content-based detection

Any defense that decides whether a message is spam by inspecting the message itself — its words, links, images, or metadata. The dominant paradigm for two decades, undone by AI-generated language that carries no tell to detect.