Skip to content

Guardrails

Runtime constraints applied to an AI system's inputs or outputs to enforce safety, format, or policy properties.

Runtime constraints applied to an AI system’s inputs or outputs to enforce safety, format, or policy properties.

Related terms


← All glossary terms