# Employer AI Interview Integrity Starter Kit Editable policy, candidate disclosure language, permission matrix, response workflow, and interview integrity technology evaluation checklist. ## How to use this template Replace bracketed text, choose rules for each interview stage, and review the final policy with recruiting, legal, privacy, security, accessibility, and technical-hiring stakeholders. This is a practical starting point, not jurisdiction-specific legal advice. --- # [Company] Candidate AI Use Policy ## Purpose We use interviews to understand your experience, judgment, communication, and ability to perform the work. This policy explains when artificial intelligence and other assistance may be used so every candidate is assessed under clear and consistent conditions. ## Scope This policy applies to [live interviews, asynchronous video responses, technical exercises, take-home work, presentations, and other listed stages]. Instructions for a specific stage take precedence when they expressly permit a tool. ## Permitted preparation You may use AI before an interview to research our organization, organize your experience, practice questions, and improve application materials, provided the information you submit remains accurate and represents your experience. ## Live interviews During a live interview, you may use only the notes, references, applications, devices, and people expressly permitted in the invitation or by the interviewer. You must answer in your own words. You may not use a tool or person to generate, recommend, or relay answers in real time unless we have told all candidates that the stage is AI-enabled. ## Technical stages The invitation will state whether documentation, internet search, autocomplete, coding copilots, or code-generation agents are allowed. You remain responsible for explaining every submission and identifying material AI assistance when disclosure is requested. ## Accessibility and language support We welcome requests for reasonable accommodation. An approved accessibility, transcription, translation, or assistive technology will not be treated as unauthorized AI use. Contact [accommodation contact] before the stage when possible; you do not need to disclose a diagnosis to an interviewer. ## Disclosure If you are uncertain whether a tool is allowed, ask [contact] before using it. If an unexpected assistive or AI feature activates, tell the interviewer promptly. Good-faith disclosure will be reviewed in context. ## Review A signal or unusual behavior will not by itself establish a violation. We may document relevant facts, ask clarifying questions, or offer a consistent reassessment. A trained person—not an automated score alone—will review material concerns before a hiring decision is made. ## Privacy Any integrity controls will be communicated as required. Information will be limited to a legitimate hiring purpose, access-controlled, retained under [retention schedule], and handled under our privacy notice and applicable law. ## Consequences Material, knowing use of prohibited assistance may result in discontinuation of the hiring process. Decisions will consider the instructions provided, evidence available, disclosed accommodations, and the candidate's explanation. ## Questions Contact [recruiting contact] for policy questions or [accommodation contact] for confidential accommodation support. --- # Permission Matrix ## Permitted - Pre-interview company or role research. - Practice questions and interview preparation. - Application or résumé editing completed before the interview. - Notes and references expressly permitted for the interview stage. ## Conditionally permitted These tools may be used only when disclosed, approved for the stage, or provided as an accommodation: - Transcription and translation. - Screen readers and other assistive technologies. - Documentation and calculators. - IDE autocomplete. - Coding copilots or code-generation tools. ## Prohibited Unless the employer has explicitly designed and announced an AI-enabled exercise: - Real-time generated answers. - Covert interview copilots. - Another person answering or relaying answers. - Remote-control assistance. - Undisclosed code generation. - Fabricated audio or video. - Recording without permission. --- # Candidate Invitation Language > For this interview, you may use your own prepared notes and any accessibility tools agreed with our recruiting team. Please do not use AI or another person to generate or suggest answers in real time. [For this technical stage, documentation is allowed; code-generation tools are not.] If you are unsure about a tool, contact [name/email] before the interview. --- # Interviewer Response Checklist ## Before the interview - [ ] Confirm the stage's measurement goal and AI-use mode. - [ ] Send the policy, stage-specific rules, and accommodation contact. - [ ] Use materially consistent instructions for comparable candidates. - [ ] Know the escalation owner and approved evidence-retention process. ## During the interview - [ ] Restate the tool rules briefly and invite questions. - [ ] Ask a job-relevant follow-up; do not accuse. - [ ] Record objective facts separately from interpretations. - [ ] Do not treat gaze, pauses, accent, nervousness, or polished wording as proof. # Suspected Unauthorized AI-Use Response Workflow 1. **Keep the conversation professional.** Do not accuse the candidate from a behavioral sign or isolated event. 2. **Clarify the work.** Ask the candidate to explain, modify, or apply the answer to a new constraint. 3. **Document objective facts.** Record the instruction, time, observable event, candidate explanation, and follow-up separately from interpretation. 4. **Check context.** Confirm which tools and accommodations were approved for the stage. 5. **Escalate for human review.** Do not make a decision from an automated probability score alone. 6. **Choose a proportionate outcome.** Close the concern, seek clarification, offer a consistent reassessment, or act when reliable evidence supports a material violation. 7. **Follow the retention rule.** Keep only approved information for the defined period. --- Starter kit source: https://capifiq.ai/blog/ai-interview-policy.html --- # Interview Integrity Technology Evaluation Checklist Use these questions after defining the policy. A written rule creates notice; the interview process still needs proportionate controls, evidence, and human review. Can your current process: - [ ] Detect hidden AI interview assistants on the interview device? - [ ] Detect suspicious remote-control tools or sessions? - [ ] Detect virtualized or manipulated interview environments? - [ ] Detect risky browser or desktop assistance relevant to the interview? - [ ] Identify technical evidence beyond ordinary screen sharing? - [ ] Preserve timestamped records for authorized review? - [ ] Support trained human review rather than automated decisions alone? - [ ] Integrate with the meeting and interview workflow your teams already use? - [ ] Support candidate notice, accommodations, privacy, and a fair review process? - [ ] State clearly which off-device methods or conditions the technology cannot verify? If the answer is no, [review your interview integrity process with Capifiq](https://capifiq.ai/get-started.html).