Human-in-the-Loop
A design paradigm where human judgment is integrated into the AI decision-making process, with humans reviewing, approving or overriding automated decisions before they take effect.
In plain language
A person reviews and approves the AI's decisions before they happen. For example, an AI might flag potentially fraudulent transactions, but a human analyst makes the final call on whether to block the transaction.
Why this matters
Human-in-the-loop is often a regulatory requirement for high-risk AI applications and a critical risk mitigation strategy. Your governance framework must define which decisions require human approval and establish standards for how humans exercise that oversight.
Relevance
GovernanceHuman-in-the-loop is a mandatory control for high-risk AI decisions and ensures human accountability in automated systems.
Related terms
Putting human-in-the-loop into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.