Online proctoring glossary

Human in the loop

Human in the loop means a person reviews AI proctoring output and makes every decision about a candidate; the software only flags.

Definition

Human in the loop is a design principle in which automated systems support people but do not make final decisions that affect individuals. In proctoring, it means AI can flag, sort and highlight evidence, but a trained person reviews it and decides whether there was a breach and what follows.

The principle has practical and legal grounds. AI analysis of video and audio is probabilistic and can be wrong, and data protection rules such as the GDPR restrict decisions based solely on automated processing that significantly affect people. The EU AI Act also requires human oversight for high-risk AI systems, which includes certain uses in education.

A real human checkpoint needs time, training, access to the full evidence and the authority to disagree with the system. It should also be visible to candidates in the exam rules. QProctor applies this principle: the AI flags and shows the evidence, and there are no automatic decisions about a candidate.

See it in practice

Try QProctor free for 14 days with up to 20 candidates.

Start free