Purpose and scope
Artificial intelligence
AI may support bounded administrative and analytical tasks, but it does not replace accountable professional judgement in assessment, recognition, complaints or appeals.
01
Permitted support
AI may assist with low-risk drafting, consistency checks, data preparation or exploratory analysis when the purpose is defined, inputs are lawfully available and a competent person verifies the output before reliance.
02
Prohibited reliance
AI must not make the final recognition, appeal, complaint or professional-award decision; invent evidence; conceal uncertainty; infer sensitive characteristics without a lawful and necessary basis; or be used where its limitations make the result unfair or unreviewable.
03
Human accountability and records
The responsible professional remains accountable for the question asked, data selected, output checked, limitations considered and decision made. Material AI use in a consequential process is documented sufficiently to permit review and challenge.
Mandatory controls
What the policy requires in operation
- Confidential or personal evidence is not entered into an unapproved AI service.
- Outputs are checked for factual error, bias, unsupported inference and data leakage.
- Automation is proportionate to the consequence and reversibility of the task.
- A person affected by a consequential decision can request meaningful human review.
Public accountability
Status changes remain visible
Where this policy affects a public accreditation or Certi Mark, the live register reflects the current status, restrictions and material public history without disclosing protected evidence or personal data.