Start with an honest disclosure
Say that the visitor is interacting with an AI assistant. Explain its intended purpose and the source boundary: for example, that it helps interpret approved careers content, rather than deciding who should be hired. Do not imply that it is a recruiter, has live knowledge of every vacancy or can make commitments on the employer’s behalf.
Show provenance and freshness
Answers need an accountable route back to their source. Show or link the approved careers page where practical, record when a source was reviewed and design a process to update or withdraw outdated content. If the assistant is uncertain, it should say so plainly and direct the candidate to an official route or person.
Design privacy into the interaction
Collect only what is necessary for the stated task. Avoid asking a candidate to disclose sensitive information merely to receive general careers guidance. The UK Information Commissioner’s Office provides guidance on data protection and AI; teams should apply the relevant law and obtain specialist advice for their deployment.
Provide a human and non-chat route
Not every question is appropriate for generated answers. A clear human escalation route, official application contact and accessible non-chat path are core service features. They are particularly important for accommodation requests, complaints, sensitive personal circumstances and questions outside the assistant’s approved knowledge.
Evaluate in the open
Set a small evaluation set of real candidate questions, test source-grounding and refusal or handoff behaviour, and review failure patterns. NIST’s AI RMF emphasises governing, mapping, measuring and managing risks. For a careers assistant, a dated evaluation and a visible limitation are more meaningful than broad claims of accuracy.
Sources
- NIST — AI Risk Management Framework: a voluntary framework for managing AI risks.
- UK ICO — Artificial intelligence and data protection: data-protection guidance for organisations using AI.
- W3C — Web Content Accessibility Guidelines (WCAG): accessibility standards relevant to candidate-facing services.