Use of AI & Automation in Assessment

Why we publish this

AIPA certifies people who build AI systems, so candidates are entitled to know how automation is used in assessing them. This policy says exactly where it is used and where it is not.

Where automation is used

Assembling exam forms: software draws each candidate's form from approved questions to match the published blueprint and rules (for example, questions that must never appear together).

Scoring: answers to multiple-choice questions are scored automatically against keys held on our servers. This is objective and the same for every candidate.

Integrity signals: your browser reports defined events — such as leaving full screen, pasting, or connecting a second display — which are logged for a reviewer. These are rule-based signals, not judgements about you.

Statistical checks: response patterns are analysed to detect possible copying or exposed questions, and question statistics are analysed to find questions that may be faulty or unfair.

Drafting support: AI tools may be used to help draft or check exam questions and course material. No question can appear on an exam until a qualified subject-matter expert has reviewed and approved it.

What is always decided by a person

Approving or rejecting your identity verification; deciding whether an attempt with integrity events is released, held or voided; any finding of misconduct and any sanction; accommodation decisions; appeals; and setting the pass mark, which is done by a panel of practitioners through a recorded standard-setting study.

We do not use facial recognition, emotion analysis, or any automated system to decide your result, your identity, or whether you cheated.

Your rights

You can ask us how any decision about you was made and challenge it through Appeals & Complaints. Where data-protection law gives you rights about automated decisions, you can exercise them through the contact details in the Privacy Notice.