AI Mock Interviewer

Abstract

Interview performance is influenced by domain knowledge, the ability to organize evidence, communication clarity, and familiarity with interview formats. Access to repeated, individualized practice can be limited by the cost and availability of human interview coaches. This paper proposes an AI Mock Interviewer, a practice-oriented system that simulates role-specific interviews and generates structured, actionable feedback. The framework combines a candidate profile and job-description parser, a competency-based question engine, text or speech input, speech-to-text processing, natural-language analysis, a transparent scoring rubric, and a feedback module. Rather than predicting employability or making hiring decisions, the proposed system is designed to support low-stakes learning. The evaluation plan specifies measures for feedback agreement with trained human raters, scoring reliability, response relevance, usability, accessibility, and fairness across user groups. A weighted rubric is presented as an initial design choice and is explicitly subject to empirical validation. Prior research on automated interview assessment indicates potential benefits in standardization and scale, while also identifying concerns about reliability, validity, bias, disability access, and applicant perceptions. Accordingly, the framework incorporates consent, data minimization, uncertainty communication, user control over recordings, and human review for consequential decisions. Since no implementation data are supplied in this paper, the contribution is a research proposal and system design, not a report of experimentally demonstrated performance.

Introduction

The text proposes an AI Mock Interviewer, a software-based platform designed to help candidates practice job interviews and improve their communication, technical knowledge, problem-solving, and response structure. Traditional mock interviews often depend on mentors, placement officers, peers, or paid coaching services, which may limit access to regular practice. The proposed system aims to provide affordable, repeatable, and personalized interview preparation through artificial intelligence.

The literature review highlights the potential of AI to assess interview responses quickly while identifying concerns about scoring reliability, fairness, accessibility, and agreement with human evaluations. Previous studies indicate that automated interview assessments may produce inconsistent results and disadvantage certain groups, including people with disabilities. Therefore, the proposed system focuses on formative learning and transparent feedback rather than automated hiring decisions.

The system allows candidates to select a target job role, experience level, interview type, response mode, and optional job description. It generates or selects relevant questions based on job-specific competencies. Candidates can respond through text or speech, with speech converted into text for analysis. Their answers are evaluated using a transparent rubric that considers relevance, technical accuracy, supporting evidence, response structure, and clarity.

The proposed system consists of several key functions:

• Adaptive question generation: Produces or selects questions aligned with the candidate's target role and experience level.

• Response processing: Analyzes text responses and transcribes speech, allowing users to correct transcription errors.

• Rubric-based assessment: Evaluates answers using a 0–4 scale across defined criteria and calculates an illustrative overall score out of 100.

• Personalized feedback: Identifies strengths, highlights missing or weak elements, suggests improved response structures, and recommends further practice.

• Progress tracking: Supports repeated practice and helps candidates identify areas requiring improvement.

The methodology follows a design-science approach involving requirement analysis, prototype development, human evaluation, and iterative improvement. The proposed evaluation will compare AI-generated assessments with ratings from trained human evaluators. It will examine scoring agreement, reliability, usability, accessibility, fairness, and changes in interview performance. Statistical analysis will include confidence intervals, effect sizes, and appropriate agreement measures.

The project also emphasizes ethical and privacy considerations, including informed consent, data minimization, secure data storage, user control over recordings and transcripts, accessibility features, bias monitoring, and transparent disclosure of AI limitations. Facial expressions, accents, emotions, and personality should not be treated as reliable indicators of professional competence without strong supporting evidence.

Conclusion

This paper presents a proposed AI Mock Interviewer framework for accessible, personalized interview practice. Its core components are role-aware question selection, text or speech response capture, response analysis, transparent rubric scoring, tailored feedback, and progress tracking. The paper emphasizes that formative coaching is a more defensible initial use than automated hiring decisions. Future work should implement a prototype, evaluate agreement with trained human raters, measure learning outcomes and usability, and conduct systematic accessibility, privacy, and fairness assessments. Claims of effectiveness should be made only after these studies are completed and their results reported.

References

[1] Hickman, L., Tay, L., & Woo, S. E. (2025). Are automated video interviews smart enough? Behavioral modes, reliability, validity, and bias of machine learning cognitive ability assessments. Journal of Applied Psychology, 110(3), 314-335. https://doi.org/10.1037/apl0001236 [2] Li, et al. (2024). Can Large Language Models Assess Personality From Asynchronous Video Interviews? A Comprehensive Evaluation of Validity, Reliability, Fairness, and Rating Patterns. IEEE Transactions on Affective Computing, 15(3), 1769-1785. https://doi.org/10.1109/TAFFC.2024.3374875 [3] Fisher, S. L., Bonaccio, S., & Connelly, C. E. (2024). AI-based tools in selection: Considering the impact on applicants with disabilities. Organizational Dynamics, 53(1), 101036. https://doi.org/10.1016/j.orgdyn.2024.101036 [4] Szanda?a, T. (2025). ChatGPT vs human expertise in the context of IT recruitment. Expert Systems with Applications, 264, 125868. https://doi.org/10.1016/j.eswa.2024.125868 [5] Biswas, S., Jung, J.-Y., Unnam, A., Yadav, G., Gupta, S., & Gadiraju, U. (2024). “Hi. I’m Molly, Your Virtual Interviewer!” Exploring the Impact of Race and Gender in AI-Powered Virtual Interview Experiences. Proceedings of the AAAI Conference on Human Computation and Crowdsourcing, 12(1), 12-22. https://doi.org/10.1609/hcomp.v12i1.31596 [6] Poe, R. L., & El Mestari, S. Z. (2024). The Conflict Between Algorithmic Fairness and Non-Discrimination: An Analysis of Fair Automated Hiring. Proceedings of the ACM conference. https://doi.org/10.1145/3630106.3659015

Copyright

Copyright © 2026 Surabhi K S, Siva A L . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.