Human accountability
A named lead reviewer is responsible for evidence selection, analysis, findings, wording and recommendations. We do not use AI to award scores, determine compliance, create accreditation or make an automated decision about an organisation or person.
Permitted, bounded assistance
Subject to the appointment and an information-risk assessment, AI tools may assist limited administrative or analytical tasks such as structuring non-sensitive notes, checking consistency or improving draft clarity. Outputs are treated as unverified and are checked against source evidence by the human reviewer.
Information controls
Client-confidential evidence is not entered into open public or consumer AI services. Personal and sensitive information is minimised. Any approved tool must have access, retention and reuse settings suitable for the material and the client's contractual requirements.
Transparency
Any material use of AI in producing an assignment deliverable is disclosed to the client. A client may ask whether AI was used, for what bounded purpose and what human verification was applied. Where the appointment prohibits AI use, that restriction is followed.
Fairness, accuracy and review
We consider confidentiality, data protection, bias, accuracy, intellectual property and explainability before use. We review this policy as tools, risks, professional expectations and regulatory guidance evolve.
Questions
Questions about AI use in a proposed or current assignment can be sent to governance@designreviewpanel.co.uk.