Fair use of AI
Updated 17 October 2026
In short: AI helps mark written and spoken answers and suggests who to look at first. It never watches video, never judges faces or voices, and never decides who gets hired. People do.
1What the AI does
- Marks written and spoken answers. The text of an answer (typed, or turned into text by the candidate's browser) goes to an AI model, which suggests a score and short notes against a published rubric.
- Writes summaries that help a recruiter read results faster.
- Suggests a shortlist. The app can list the people above a pass mark you choose. It is a list of suggestions only.
- Everything else is plain rules: multiple-choice answers, games and technical tasks are marked by fixed answer keys, not by AI.
2What it never does
- It does not watch candidate video and does not analyse faces, voices, accents, emotions or body language.
- It does not recognise or identify faces. The "camera covered" check only looks at whether the picture is dark or blank.
- It does not reject anyone, email anyone or move anyone forward by itself.
- It does not use race, gender, age, religion, disability, nationality or similar traits, and none of these are asked for in an assessment.
3People decide
Scores and AI notes are a guide for the recruiter. A recruiter can read the actual answer, change an AI rating, and add notes. The suggested shortlist waits for a person to approve each name. Hiring decisions are made by people, never by the AI alone.
4What candidates are told
Candidates see what is recorded and how long recordings are kept before a test starts, and our Privacy Policy says that written and spoken answers are marked with the help of AI. Where the law requires it, the employer should also tell candidates in the invitation. Candidates can ask the recruiter, or us, for a person to review their result.
5Fairness and the law
Rules about automated hiring tools differ by place, for example New York City Local Law 144 (bias audits and notices), the EU AI Act (hiring tools are treated as high risk) and Illinois rules on AI in video interviews. The employer using a tool is usually responsible for meeting them. We design HireVation so that a human stays in charge, but you remain responsible for checking what applies to your hiring and for telling candidates when the law requires it. Please take legal advice if you are unsure.
6What you can check
- Every score can be traced to the answers behind it.
- Skills-lab interviews and exercises are rated on a published 1 to 5 scale, and a recruiter's rating replaces the AI's.
- The same questions are asked of everyone who takes the same assessment.
7Honest limits
AI can be wrong. Spoken answers depend on how well the browser turns speech into text, which can be worse for some accents, microphones and noisy rooms. HireVation has not yet had an independent bias audit. Use scores to compare and to decide who to talk to, not as the only reason to exclude someone. Questions or concerns: support@hirevation.com.
