Maintaining Assessment Integrity in Video Interviews
Learn how to maintain the integrity of video interviews through effective question design, candidate expectation-setting, and built-in platform tools that help identify overly prepared or AI-assisted responses.
Maintaining Assessment Integrity in Video Interviews
Ensuring the integrity of a video interview is a common concern, particularly as candidates gain access to AI tools and shared preparation materials.
The good news is that with the right design choices and platform features, you can significantly reduce the risk of over-prepared or AI-assisted responses — while still delivering a fair, candidate-friendly experience.
This guide outlines the most effective tactics and how to prioritise them.
Why Integrity Risks Occur
Video interviews are most vulnerable when:
- Questions are predictable or easily searchable
- Candidates have extended time to prepare responses
- There are no signals or deterrents against AI-assisted answers
Addressing these areas directly is the key to maintaining authenticity.
Core Principles for Strong Assessment Design
1. Set Expectations Early
Start by clearly communicating expectations to candidates.
Use your welcome lobby to:
- Reinforce that responses should reflect their own thoughts and experiences
- Clarify your stance on AI usage
- Reference any Honesty Contract or assessment guidelines
This simple step acts as a powerful behavioural nudge and reduces misuse before it starts.
2. Design Questions That Reward Authenticity
The question set is the single biggest driver of integrity.
Focus on:
Reflective / Experience-Based Questions
- Ask candidates to draw on real experiences
- Example: “Tell us about a time you handled a difficult stakeholder…”
These are harder to fabricate convincingly and less suited to AI-generated answers.
Applied / Contextual Questions
- Tailor questions to your organisation or role
- Example: “Which of our values resonates most with you and why?”
3. Use Question Variants
Question variants reduce the risk of answer sharing between candidates.
- Rotate multiple versions of the same question
- Ensure consistency in what’s being assessed, not the wording
This is especially valuable in high-volume processes (e.g. admissions, early careers).
4. Reduce Over-Preparation Opportunities
Shorter thinking time helps limit the ability to:
- Look up answers
- Generate responses using AI tools
This encourages more natural, in-the-moment answers that better reflect the candidate.
5. Introduce Real-World, Hard-to-Fake Tasks
As an additional guard (or for higher-stakes roles), incorporate:
Screen-sharing or recording-based questions, for example:
- “Open our website and talk us through a value that resonates with you”
- “Share an example of recent work and explain your approach”
These:
- Increase authenticity
- Require real-time engagement
- Are significantly harder to script or outsource
6. Leverage Built-In Platform Signals
Shortlister includes features designed to highlight potential integrity concerns:
Word / Error Count Analysis
- Flags responses that may be overly scripted or unnatural
Authenticity Indicators
- Surface signals that may suggest AI-assisted or heavily prepared answers
These are not designed to penalise candidates, but to support reviewers in making informed decisions.
Recommended Approach
To balance integrity with candidate experience, we recommend:
Phase 1 (Essential)
- Set clear expectations in the welcome lobby
- Strengthen your question design (reflective, contextual)
- Introduce question variants
- Reduce thinking time where appropriate
Phase 2 (Enhanced Integrity)
- Add screen-sharing or task-based questions
- Use platform indicators to support reviewer judgement
Final Thought
Maintaining integrity isn’t about eliminating support tools entirely, it’s about designing an experience where authentic responses are the easiest and most natural option for candidates.
With the right combination of expectation-setting, question design, and platform features, video interviews remain a highly effective and fair assessment method.