Write a Candidate Policy on Using AI in Your Process

When to use it

The prompt

Draft a short, candidate-facing policy on using AI tools during our hiring process. OUR STAGES: {{e.g. application, recruiter screen, take-home, live technical interview, final round}} OUR POSITION: {{what we're comfortable with, e.g. AI fine for resumes and take-homes if disclosed, not in live interviews; or AI allowed throughout because our engineers use it daily}} HOW OUR TEAM WORKS: {{e.g. engineers use coding agents daily; or AI tools are restricted for security reasons}} TONE: {{e.g. warm and plain-spoken}} Write: 1. A short opening explaining why we have a policy, written as helping candidates rather than warning them. 2. For each stage, what's fine, what isn't, and whether we'd like it mentioned. Where AI is allowed, say how we'll assess the candidate's own judgement rather than the tool's output. 3. For a live coding or problem-solving interview, whether they may use an AI assistant, and if so, that we'll ask them to talk through what it produced and what they'd change. 4. What we do with AI on our side, such as using it to help schedule or summarise notes, and a clear statement that a human makes every hiring decision (only if that's true; ask me if you're unsure). 5. How to ask for adjustments or raise a question. Then separately: 6. Guidance for our interviewers on applying the policy consistently, including not penalising someone for using AI where the policy allows it. 7. Anything in my position that contradicts how the team really works, since candidates will notice. Under 300 words for the candidate-facing part. Plain language, no legal jargon.

Tip

Match the policy to how the team really works. If your engineers use coding agents every day, banning them in the take-home tests a skill the job doesn't need. Allowing them and asking the candidate to defend the output tests the one it does.

How to use it

Paste the prompt into ChatGPT, Claude or whichever assistant you use, then replace every {{bracketed}} part with your own detail. The more specific and messier your input, the better the output — a model given raw notes has more to work with than one given a tidy summary you wrote first.

More Interviewing prompts

Open the full tool for all 87 prompts, the glossary, and the JD decoder.