The AI life cycle is the full path of an artificial intelligence system, from initial design and data selection through deployment, monitoring, updates, and eventual retirement. It treats AI as a living service, not a one-time product. Each stage can change how the system behaves, who controls it, and what risks it introduces.
In cyber security, the life cycle matters because attacks and failures can happen at any stage. Adversaries may try to poison training data, abuse poorly reviewed updates, or exploit weak access controls during deployment. Defenders use the same life cycle view to reduce risk: approve changes, log outputs, monitor drift, test for unsafe behavior, and retire systems that can no longer be trusted. For public-facing tools like chatbots, life-cycle governance also supports transparency, accountability, and incident response.