Abrir menu de navegação
AIDive
PT
Entrar
Voltar ao glossário

AI Model Governance

Ethics & Safety

Model control processes throughout the entire life cycle: from selection and training to launch, monitoring, updating and decommissioning.

Definição

Managing AI models helps you stay in control of which model is running, what data it was trained on, who approved it, what risks are known, and when it needs to be updated. This is especially important when the company has many models, versions and teams.

Exemplo

After updating the support model, the team commits the version, compares metrics, checks security, and stores information about the reasons for the transition.

Por que importa

The term is important for production AI: a model is not a one-time file, but a system that needs to be maintained.

Como funciona

Management includes model registries, documentation, versions, accesses, metrics, monitoring, auditing, responsible roles and change procedures.

Onde é usado

  • MLOps
  • corporate AI systems
  • model version control

Limitações

Without discipline, it's easy to launch the wrong version, lose reproducibility, or not notice the degradation in quality after data changes.

FAQ

Why is “AI Model Governance” useful to know?

The term is important for production AI: a model is not a one-time file, but a system that needs to be maintained.