A generative system is software that creates new text, images, code, or other outputs from a prompt. It does not simply retrieve stored content; it predicts and assembles a fresh response based on patterns learned during training. In security terms, that makes the system powerful but also hard to control.
Generative systems matter because they can be used to scale both defense and abuse. Defenders use them for triage, summarization, and support automation, but they must add safeguards such as access controls, output filtering, logging, and human review. Attackers can use the same tools to draft phishing messages, generate malicious code, impersonate people, or produce misleading content at high volume. Security teams also watch for prompt injection and unsafe output, where crafted inputs push the system to reveal sensitive data or bypass policy. Managing a generative system is therefore a core part of modern platform and application security.