Human Resource Management

Who Gets the Credit? Generative AI Contribution Disclosure and Performance Appraisal Fairness among Faculty Members in Indian Higher Education Institutions

Ankur Khandelwal

Department of Management Studies, India

Published November 2026

Abstract

Generative artificial intelligence (GenAI) is increasingly used by knowledge workers to prepare reports, teaching materials, research summaries, presentations and administrative documents. Although such tools may improve speed and output quality, they create a human resource management question that is not resolved by measuring productivity alone: how should an employee receive performance credit when a work product is created through a combination of human judgement and AI assistance? This conceptual paper develops a research framework to examine the relationship between GenAI contribution disclosure, perceived ownership of work, procedural fairness in performance appraisal and employees’ trust in institutional HR practices among faculty members in Indian higher education institutions. It proposes that transparent disclosure of the type and extent of AI assistance may improve appraisal fairness when institutions use clear contribution criteria and retain human review. However, disclosure may also create a perceived competence penalty if AI-assisted work is interpreted as evidence of lower individual capability. The paper proposes a quantitative, multi-source research design involving faculty members and their academic supervisors. It does not report empirical findings; the hypotheses and framework are intended for subsequent field testing. The study seeks to extend HRM research by connecting human–AI contribution attribution with faculty performance management in the Indian higher education context.

Keywords: Generative artificial intelligence; performance appraisal; contribution attribution; perceived work ownership; procedural justice; higher education; human resource management

1Introduction

Generative artificial intelligence has moved from an experimental technology to a practical workplace assistant. Employees can use it to structure reports, summarise information, draft communication, generate teaching resources and support analysis. In higher education, faculty members may use these tools for lesson planning, assessment design, research support and institutional documentation. The resulting work is often neither entirely human-produced nor entirely machine-produced.

This creates a new HR question. Traditional performance appraisal systems generally assess outputs, competencies, behaviours and goal achievement. When an employee submits an AI-assisted output, the organisation must decide what proportion of the result reflects the employee’s own knowledge, judgement, verification and effort. If appraisal systems reward only the final output, employees who use AI effectively may receive credit without transparent recognition of the process. Conversely, if AI use is treated as a sign of reduced capability, employees may be penalised despite having applied substantial professional judgement.

Recent scholarship has begun to examine AI attribution and workplace disclosure. Research has explored how people assign credit in human–AI co-creation [1], how attribution of identical work to AI or a human changes perceived task meaning and effort [2], and how disclosure practices shape stigma and willingness to reveal AI use [3]. Research has also examined AI involvement in performance appraisal [4]. These studies establish that attribution and appraisal are active research areas.

The proposed contribution is narrower: examining how institutions translate disclosed levels of GenAI assistance into faculty performance credit, perceived procedural fairness and trust in appraisal, within Indian higher education institutions. The context-specific focus, the distinction between types of AI contribution, and the proposed connection to appraisal policy provide a testable research opportunity.

2Research problem and gap

Existing research on human–AI collaboration has addressed questions of authorship, credit allocation, disclosure stigma and evaluation of AI-supported work [1–4]. However, the literature indicates an opportunity to examine how these issues are operationalised in faculty appraisal systems in Indian higher education. In particular, the proposed study distinguishes between AI use for language correction, idea generation, content drafting, data analysis and decision support. These activities involve different levels of human intellectual contribution and should not automatically be treated as equivalent.

The central research problem is whether faculty members believe that institutional appraisal systems fairly recognise their own contribution when GenAI has supported the production of work. A related issue is whether a clear disclosure policy reduces uncertainty or instead creates fear that AI-assisted work will be valued less.

3Objectives of the study

  1. 1. To examine the relationship between the extent and type of GenAI assistance disclosure and perceived fairness of faculty performance appraisal.
  2. 2. To assess whether perceived ownership of AI-assisted work influences employees’ perceptions of performance credit.
  3. 3. To investigate whether perceived competence penalty mediates the relationship between AI disclosure and appraisal fairness.
  4. 4. To examine the role of institutional AI-use policy clarity in shaping disclosure and trust in appraisal.
  5. 5. To propose HR policy recommendations for fair recognition of human contribution in AI-assisted academic work.

4Research questions

  1. 1. Does disclosure of GenAI assistance influence faculty members’ perceptions of performance appraisal fairness?
  2. 2. Do faculty members perceive different levels of ownership for AI-assisted work depending on the type of AI contribution?
  3. 3. Does fear of a competence penalty reduce willingness to disclose AI assistance?
  4. 4. Does clarity of institutional AI-use policy strengthen the relationship between disclosure and appraisal fairness?
  5. 5. Does perceived appraisal fairness contribute to trust in institutional HR practices?

5Conceptual background and theoretical foundation

5.1Attribution theory

Attribution theory explains how individuals interpret the causes of outcomes and allocate responsibility or credit. In AI-assisted work, supervisors may attribute a successful output to employee expertise, AI capability or a combination of both. The proposed study applies this perspective to performance credit: the employee’s perceived share of contribution may depend on the nature of the task, the degree of AI assistance and the quality of human verification.

5.2Organisational justice

Organisational justice concerns employees’ perceptions of fairness in decision-making procedures and outcomes. For this study, procedural justice is relevant because faculty members need to understand how AI-assisted work is evaluated, what evidence is considered and whether they can explain or challenge an appraisal decision. A transparent and consistently applied contribution rubric may improve perceived fairness.

5.3Psychological ownership

Psychological ownership refers to the feeling that a person’s work or output is ‘mine’. When AI generates substantial portions of an output, the employee may experience reduced ownership, even when the employee selects prompts, evaluates responses, corrects errors and takes responsibility for the final work. This study treats perceived ownership as a possible explanatory variable rather than assuming that AI use necessarily reduces ownership.

6Proposed conceptual framework

The proposed model places GenAI contribution disclosure as the focal predictor. Perceived work ownership and perceived competence penalty are proposed as explanatory mechanisms. Institutional AI policy clarity is proposed as a moderator, while perceived performance appraisal fairness and trust in institutional HR practices are the principal outcomes.

The proposed relationships are as follows: GenAI contribution disclosure → perceived work ownership → appraisal fairness; GenAI contribution disclosure → perceived competence penalty → appraisal fairness; institutional AI policy clarity moderates the disclosure–fairness relationship; appraisal fairness → trust in institutional HR practices.

7Hypotheses

H1: Clearer disclosure of the type and extent of GenAI assistance is positively associated with perceived performance appraisal fairness.

H2: Perceived work ownership is positively associated with perceived performance appraisal fairness.

H3: Perceived competence penalty is negatively associated with perceived performance appraisal fairness.

H4: Perceived work ownership mediates the relationship between GenAI contribution disclosure and appraisal fairness.

H5: Perceived competence penalty mediates the relationship between GenAI contribution disclosure and appraisal fairness.

H6: Institutional AI-use policy clarity strengthens the positive relationship between GenAI contribution disclosure and perceived appraisal fairness.

H7: Perceived performance appraisal fairness is positively associated with trust in institutional HR practices.

8Research methodology

8.1Research design

A quantitative, cross-sectional explanatory design is proposed, with a short qualitative pilot to refine the questionnaire. A stronger design would collect employee responses and supervisor appraisal assessments at two time points to reduce common-method bias.

8.2Population and scope

The proposed population is faculty members working in recognised higher education institutions in Madhya Pradesh, India. The study may include public and private universities and affiliated colleges, subject to access and institutional permission. Respondents should have used at least one GenAI tool for a work-related academic or administrative task during the previous six months.

8.3Sample and sampling

A target of approximately 300 faculty respondents is proposed for the main survey, subject to a formal power analysis and the number of predictors in the final model. A stratified approach may be used to include institutions by ownership type, discipline and designation. A separate supervisor sample may be collected where feasible.

8.4Data collection instrument

A structured questionnaire may use a five-point Likert scale ranging from strongly disagree to strongly agree. Items should be adapted from established scales for organisational justice, psychological ownership, trust and perceived competence threat, then piloted with faculty members. New items on AI contribution disclosure should be developed through expert review and cognitive interviews.

8.5Suggested variables and operationalisation

The variables of the proposed model and their possible measurement focus are summarised in Table 1.

Table 1. Suggested variables and operationalisation.
Variable Meaning in this study Possible measurement focus
GenAI contribution disclosure How clearly an employee reports AI involvement Type of task, extent of AI contribution, human review
Perceived work ownership Feeling of personal ownership over the final output Responsibility, control, personal intellectual contribution
Perceived competence penalty Expectation of being judged as less capable because AI was used Fear of negative evaluation, perceived skill discount
AI policy clarity Clarity of institutional rules for acceptable AI use and disclosure Written guidance, examples, consistent communication
Appraisal fairness Perceived fairness of evaluation procedures and credit allocation Consistency, voice, explanation, appeal opportunity
Trust in HR practices Confidence in the fairness and reliability of institutional appraisal Reliability, transparency, confidence in HR decisions

8.6Proposed statistical techniques

Data may be analysed using descriptive statistics, reliability analysis, confirmatory factor analysis and structural equation modelling. Mediation and moderation may be tested using bootstrapped indirect effects or an appropriate SEM procedure. Group comparisons may examine differences by designation, discipline, institution type and frequency of GenAI use. The final analysis plan should be selected after scale validation and assessment of sample adequacy.

9Ethical considerations

Participation should be voluntary and based on informed consent. No individual appraisal score or identifiable employment record should be collected without explicit institutional approval. Responses should be anonymised, stored securely and reported only in aggregate. The study should not encourage disclosure of confidential student, research-participant or institutional information to public AI tools.

10Expected contributions and HR implications

The study is expected to offer a practical way to distinguish AI assistance from the employee’s substantive contribution. It may help institutions move beyond blanket rules that either prohibit AI use or reward outputs without considering how they were produced. A contribution rubric could separately recognise problem definition, prompt design, subject expertise, verification, contextual adaptation, originality and accountability.

For HR departments, the proposed framework may support clearer appraisal guidelines, supervisor training and an employee right to explain AI-assisted work. Institutions could specify which uses require disclosure, how disclosure will be assessed and how human judgement will be retained. These are proposed implications to be evaluated after data collection, not empirical conclusions.

11Limitations and future research

The proposed study is limited to higher education institutions in one Indian state and may not generalise to other sectors or regions. Self-reported perceptions may differ from actual appraisal decisions. Cross-sectional data cannot establish causality. Future research could use vignette experiments in which supervisors evaluate identical outputs with different disclosure statements, or longitudinal data linking disclosure practices to actual appraisal outcomes.

12Conclusion

Generative AI changes not only how work is completed but also how individual contribution may be recognised. For HRM, the key issue is not simply whether employees use AI; it is whether appraisal systems can fairly distinguish technological assistance from human expertise, judgement and accountability. This paper proposes a research framework focused on faculty members in Indian higher education institutions. The framework should now be tested using primary data; no empirical results are claimed in this conceptual paper.

References

  1. [1]He J, Houde S, Weisz JD. Which contributions deserve credit? Perceptions of attribution in human-AI co-creation. In: Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI ’25); 2025; Yokohama, Japan. p. 1–18. https://doi.org/10.1145/3706598.3713522
  2. [2]Nikolova M, Milanova V, Wang F. When AI does the work: Does attribution shape meaning and effort? IZA Discussion Paper No. 18784. Bonn: IZA Institute of Labor Economics; 2026. https://www.iza.org/publications/dp/18784/when-ai-does-the-work-does-attribution-shape-meaning-and-effort
  3. [3]He J, Akella A, Do HJ, Weisz J. Overcoming barriers to AI disclosure: The role of granularity and sociocultural enablers in reducing perceived stigmas. In: ACM Conference on Fairness, Accountability, and Transparency (FAccT 2026); 2026. https://research.ibm.com/publications/overcoming-barriers-to-ai-disclosure-the-role-of-granularity-and-sociocultural-enablers-in-reducing-perceived-stigmas
  4. [4]Pan Y, Froese FJ, Xue S. The role of AI in performance appraisal: A mixed-method study of employee experience through a relational lens. Hum Resour Manage. 2026;65(3):781–798. https://doi.org/10.1002/hrm.70049
  5. [5]Khosrowi D, Finn F, Clark E. Engaging the many-hands problem of generative-AI outputs: A framework for attributing credit. AI Ethics. 2025;5(5):4495–4513. https://doi.org/10.1007/s43681-024-00440-7
  6. [6]Edmondson A. Psychological safety and learning behavior in work teams. Adm Sci Q. 1999;44(2):350–383. https://doi.org/10.2307/2666999
  7. [7]Greenberg J. A taxonomy of organizational justice theories. Acad Manage Rev. 1987;12(1):9–22. https://doi.org/10.5465/amr.1987.4306437