Project – Generative AI Use, Information Verification Behaviour and Research Integrity Among Postgraduate Students: A Study of Selected Nigerian Universities.

CHAPTER ONE

INTRODUCTION

Title: Generative AI Use, Information Verification Behaviour and Research Integrity Among Postgraduate Students: A Study of Selected Nigerian Universities

1.1 Background to the Study

The rapid development of artificial intelligence has introduced significant changes into the ways information is produced, accessed, processed and communicated across different sectors of society. One of the most influential developments in this regard is Generative Artificial Intelligence (GenAI), which refers to artificial intelligence systems capable of producing new content, including written texts, summaries, images, computer codes, explanations and other forms of information in response to users’ instructions. Unlike conventional digital tools that mainly retrieve or organise existing information, generative AI systems use patterns learned from large datasets to generate responses that resemble human-produced content. The emergence of applications such as ChatGPT, Google Gemini, Microsoft Copilot, Claude and other large language model-based systems has expanded the possibilities for information access and knowledge production. In higher education, these tools are increasingly being considered for academic writing, research assistance, literature exploration, language improvement and problem-solving. Miao and Holmes (2023) observed that generative AI offers opportunities for transforming teaching, learning and research, while also raising important concerns relating to accuracy, privacy, human agency and ethical responsibility. Consequently, the growing presence of generative AI in universities has created a need to understand not only how students use these technologies but also how they evaluate the information generated and maintain acceptable standards of academic and research conduct.

The postgraduate level of university education places considerable demands on students because they are expected to demonstrate advanced knowledge, independent thinking, critical analysis and the ability to contribute meaningfully to existing scholarship. Postgraduate students are required to identify research problems, review relevant literature, develop theoretical and conceptual arguments, select appropriate methodologies, analyse evidence and communicate findings in a scholarly manner. These activities often involve extensive engagement with academic databases, journals, books, institutional repositories and other information resources. Generative AI tools have become attractive within this research environment because they can assist students in explaining complex concepts, generating research ideas, summarising lengthy materials, improving grammar, organising arguments and identifying possible directions for further investigation. However, the usefulness of these tools depends significantly on how students employ them. Yakubu, Ibrahim and Digma (2024), in their study of postgraduate students at Sa’adu Zungur University, Bauchi State, examined the factors influencing the adoption of ChatGPT for academic support and enhanced learning. Their study demonstrates that generative AI adoption is already a subject of empirical investigation among Nigerian postgraduate students. Similarly, English, Nash and Mackenzie (2025) examined the use and understanding of generative AI among postgraduate researchers in the United Kingdom and highlighted the importance of examining how advanced students incorporate these technologies into their research practices. These developments suggest that postgraduate students should not be viewed merely as users of new digital applications but as researchers whose engagement with generative AI may influence the quality, originality and credibility of their academic work.

One important advantage of generative AI is its potential to improve the efficiency of academic information processing. Students may use these tools to obtain explanations of unfamiliar theories, develop preliminary outlines, identify relevant keywords, simplify technical language, generate questions for literature exploration and improve the presentation of research documents. Generative AI can also support students who experience difficulties with academic language, particularly when organising complex ideas into coherent scholarly expressions. In this regard, the technology may contribute to improved accessibility, productivity and learning support when used appropriately. Yusuf, Pervin and Román-González (2024), in a multinational study involving participants from 76 countries, found that generative AI was increasingly being used in higher education, particularly for information retrieval and text paraphrasing. Their findings also identified concerns relating to academic dishonesty and the need for ethical guidelines. Johnston et al. (2024), in a study involving 2,555 university students, similarly reported substantial student awareness and use or consideration of generative AI for academic purposes, while identifying different levels of acceptance depending on the nature of the task. These findings demonstrate that generative AI use is not limited to a particular academic discipline or geographical location. Nevertheless, the benefits of these tools should be understood in relation to the responsibilities of postgraduate research, because the ability to generate information quickly does not automatically establish that the information is accurate, relevant, original or suitable for academic citation.

A major concern associated with generative AI use is the reliability of information produced by these systems. Although generative AI applications can provide fluent, detailed and apparently authoritative responses, they may also produce inaccurate explanations, outdated information, fabricated references, incorrect statistics and statements that cannot be supported by credible evidence. This problem is commonly discussed in the literature as AI hallucination, referring to instances in which an AI system generates information that appears convincing but is false, misleading or unsupported. Rawte et al. (2023) examined the emergence of hallucination in large language models and explained the implications of such errors for the reliability of AI-generated content. McKenna et al. (2023) also demonstrated that language models can produce incorrect inferences because of biases and patterns learned during pre-training. More recent work by Huang et al. (2025) further discusses the distinction between fluent generated responses and information that can be substantiated through reliable evidence. These limitations are particularly important in postgraduate research, where inaccurate information can affect the formulation of research problems, interpretation of scholarly arguments, selection of theoretical perspectives and presentation of findings. A student who accepts AI-generated information without examining its accuracy may unknowingly introduce false claims, misleading interpretations or non-existent sources into a dissertation or thesis. Therefore, the growing use of generative AI makes information verification an essential part of responsible postgraduate research.

Information verification behaviour refers to the deliberate actions individuals take to determine whether information is accurate, credible, relevant, current and supported by reliable evidence. In academic research, verification involves checking AI-generated claims against original journal articles, scholarly books, official publications, research reports, recognised databases and other dependable sources. It also involves confirming authors’ names, publication dates, journal titles, volume and issue details, page numbers, digital object identifiers and the actual existence of cited publications. Verification may further require comparing information from multiple independent sources, examining the context in which evidence was produced and distinguishing between established facts, interpretations and unsupported opinions. UNESCO’s guidance on generative AI in education and research emphasises the need for human-centred, ethical and responsible integration of these technologies, including attention to the limitations of AI-generated information (Miao & Holmes, 2023). In the same direction, the Committee on Publication Ethics (COPE Council, 2023) maintains that researchers remain responsible for the accuracy and integrity of work produced with AI assistance and that the use of such tools should be transparently disclosed where applicable. These principles indicate that postgraduate students should not treat generative AI as an independent authority whose responses require no further examination. Rather, they should use it as a support tool while retaining responsibility for checking, interpreting and substantiating the information incorporated into their research.

The issue of information verification is closely connected with research integrity, which represents the commitment to honesty, transparency, accountability, accuracy and responsible conduct throughout the research process. Research integrity requires postgraduate students to present information truthfully, acknowledge the contributions of other scholars, avoid plagiarism, maintain accurate records, report research procedures honestly and refrain from fabricating or falsifying evidence. The use of generative AI introduces additional ethical questions concerning authorship, originality, acknowledgement, data confidentiality and the extent to which AI-generated materials may be incorporated into academic work. For example, a student who submits AI-generated arguments as entirely independent intellectual work without following institutional disclosure requirements may undermine transparency and authorship expectations. Similarly, the inclusion of fabricated references or unverified claims can compromise the credibility of a research document even where the student did not deliberately intend to mislead. COPE Council (2023) explains that AI systems cannot be treated as authors because they cannot accept responsibility for the work, manage conflicts of interest or take accountability for ethical and legal obligations. The Committee further stresses that human authors remain responsible for all aspects of manuscripts produced with AI assistance. Gonsalves (2025) also identifies non-compliance with AI-use declarations as a significant challenge for academic integrity, linking it to unclear guidelines, fear of academic consequences and inconsistent institutional practices. These concerns make it necessary to examine whether postgraduate students’ use of generative AI is accompanied by responsible verification and transparent research practices.

In Nigeria, the growing availability of digital technologies and generative AI applications has created new opportunities and challenges for universities, postgraduate schools, supervisors, librarians and research administrators. Nigerian postgraduate students operate within diverse institutional environments characterised by differences in digital infrastructure, access to scholarly databases, research funding, supervisory arrangements, information literacy training and institutional guidelines. These conditions may influence the extent to which students adopt generative AI, the purposes for which they use it and their ability to verify the information it produces. Yakubu et al. (2024) provide relevant Nigerian evidence by examining ChatGPT adoption for academic support among postgraduate students at Sa’adu Zungur University. Their study establishes a local basis for investigating postgraduate engagement with generative AI, although the broader relationship between AI use, information verification behaviour and research integrity requires further attention. The international evidence also indicates that student use of generative AI is accompanied by concerns about academic honesty, institutional regulation and responsible use (Johnston et al., 2024; Yusuf et al., 2024). However, the experiences of postgraduate students in selected Nigerian universities cannot automatically be assumed to be identical to those reported in other countries or among undergraduate populations. It is therefore important to investigate the issue within the Nigerian university environment, taking account of students’ patterns of AI use, their verification practices and the extent to which these behaviours support or threaten research integrity. This study is consequently designed to examine generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities.

1.2 Statement of the Problem

The increasing availability of generative AI applications has changed the way postgraduate students search for information, develop academic arguments and undertake research-related activities. Tools such as ChatGPT, Gemini, Copilot and Claude can provide explanations, generate ideas, summarise materials and assist with academic writing within a relatively short period. Although these capabilities may improve research efficiency, they also create concerns about students’ dependence on AI-generated information and the possible reduction of independent intellectual engagement. Postgraduate research is expected to reflect the student’s ability to identify credible evidence, interpret scholarly arguments and make original contributions to knowledge. However, when students rely excessively on generative AI to develop research ideas, construct arguments or prepare substantial sections of academic documents, there is a possibility that the research process may become more dependent on automated responses than on independent scholarly investigation. Yakubu et al. (2024) have established that postgraduate students in a Nigerian university are adopting ChatGPT for academic support and enhanced learning, while Yusuf et al. (2024) identified both the educational benefits and integrity concerns associated with generative AI use in higher education. The problem, therefore, is not simply that postgraduate students use generative AI, but whether their use is informed, controlled and consistent with the intellectual responsibilities required of postgraduate research.

Another major problem concerns the verification of information generated by artificial intelligence systems. Generative AI tools can produce responses that appear convincing and academically appropriate, even when some of their contents are inaccurate, incomplete or unsupported by credible sources. This creates a serious challenge for postgraduate students who may use such responses in literature reviews, theoretical discussions, research backgrounds and other sections of their dissertations or theses. A student who accepts AI-generated information without checking the original sources may unknowingly include fabricated references, incorrect publication details, misleading interpretations or unsupported claims in academic work. Rawte et al. (2023) identified hallucination as a significant limitation of large language models, while McKenna et al. (2023) demonstrated how language models may produce incorrect inferences because of learned patterns and biases. Despite these limitations, the extent to which postgraduate students consistently verify AI-generated information through academic databases, original publications, official documents and independent sources remains an important issue requiring investigation. In particular, it is necessary to establish whether students distinguish between information generated by AI and information supported by verifiable scholarly evidence.

The situation also raises important questions about research integrity among postgraduate students. Research integrity requires honesty in the production and presentation of academic knowledge, proper acknowledgement of sources, originality, transparency and accountability for research processes and outcomes. The use of generative AI may create ethical difficulties where students present AI-generated content without appropriate acknowledgement, submit unverified references, reproduce inaccurate information or allow AI applications to perform intellectual tasks that should remain their responsibility. COPE Council (2023) emphasises that AI tools cannot take responsibility for scholarly publications and that human researchers remain accountable for the accuracy and integrity of work produced with AI assistance. Similarly, Gonsalves (2025) identifies failure to comply with AI-use declaration requirements as a challenge to transparency and academic integrity in higher education. However, the relationship between postgraduate students’ level of generative AI use, their information verification behaviour and their commitment to research integrity requires further empirical examination within Nigerian universities. Without sufficient understanding of these relationships, universities may find it difficult to develop effective policies, training programmes and supervisory guidelines that encourage responsible AI use without unnecessarily discouraging its legitimate academic benefits.

Although studies have examined generative AI adoption, student perceptions and academic integrity in different higher education settings, there remains a need for more focused evidence concerning the combined relationship between generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities. For instance, Yakubu et al. (2024) examined factors influencing ChatGPT adoption for academic support among postgraduate students, but the present study extends the discussion by considering how AI use relates to the verification of information and the maintenance of research integrity. The central concern is whether postgraduate students who frequently use generative AI also demonstrate adequate verification practices and responsible research conduct, or whether uncritical dependence on these tools creates risks for the credibility and ethical quality of their academic work. Differences in institutional policies, access to research resources, digital literacy and awareness of acceptable AI practices may also influence these behaviours. It is against this background that the study seeks to examine generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities.

1.3 Aim and Objectives of the Study

The main aim of this study is to examine generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities.

The specific objectives are to:

1. Examine the extent and patterns of generative AI use among postgraduate students in selected Nigerian universities.

2. Assess the information verification behaviours adopted by postgraduate students when using generative AI for academic and research activities.

3. Determine the relationship between generative AI use and information verification behaviour among postgraduate students in selected Nigerian universities.

4. Examine the relationship between generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities.

1.4 Research Questions

The following research questions will guide the study:

1. What is the extent and pattern of generative AI use among postgraduate students in selected Nigerian universities?

2. What information verification behaviours do postgraduate students adopt when using generative AI for academic and research activities?

3. What is the relationship between generative AI use and information verification behaviour among postgraduate students in selected Nigerian universities?

4. What is the relationship between generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities?

1.5 Research Hypothesis

The following null hypothesis is formulated to guide the study and will be tested at a 0.05 level of significance:

H₀₁: There is no statistically significant relationship between generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities.

1.6 Significance of the Study

The findings of this study are expected to be useful to postgraduate students, university management, postgraduate schools, academic supervisors, university librarians, research administrators and future researchers.

Postgraduate Students: The study will help postgraduate students understand the importance of responsible generative AI use in academic research. It will draw attention to the need to verify AI-generated information before incorporating it into research documents. The findings may also encourage students to distinguish between using AI as a research-support tool and depending on it to replace independent thinking, scholarly reading and critical analysis.

University Management: The findings will provide university management with information that may assist in developing institutional policies and guidelines on the acceptable use of generative AI in teaching, learning and research. The study may help universities establish clear expectations concerning AI-use disclosure, academic authorship, originality, information verification and research accountability.

Postgraduate Schools: The study will be useful to postgraduate schools in designing training programmes and orientation activities that address the ethical and academic implications of generative AI. It may encourage the inclusion of AI literacy, research ethics and information verification in postgraduate research training and research methodology courses.

Academic Supervisors: The findings may assist supervisors in understanding the ways postgraduate students use generative AI during proposal development, literature review, data analysis and dissertation writing. It may also help supervisors develop more effective approaches to discussing AI use with their students, checking the credibility of research sources and encouraging independent academic judgement.

University Librarians: The study will provide useful insights for academic librarians who support students in identifying, accessing and evaluating scholarly information. Its findings may strengthen the development of information literacy programmes that teach postgraduate students how to verify AI-generated references, locate original publications and assess the credibility of digital information.

Research Administrators and Ethics Committees: The findings may support research administrators and institutional ethics committees in reviewing existing guidelines on research integrity in the context of generative AI. The study may also provide a basis for improving institutional approaches to transparency, authorship, originality and accountability in postgraduate research.

Future Researchers: The study will serve as a source of information for researchers interested in artificial intelligence, higher education, information literacy, research ethics and academic integrity. It may also provide a foundation for further studies examining generative AI use across different academic disciplines, postgraduate levels and university environments.

1.7 Scope of the Study

The study focuses on generative AI use, information verification behaviour and research integrity among postgraduate students in selected Nigerian universities. Geographically, it will be limited to the universities selected for the research. The specific institutions will be identified based on the researcher’s approved study locations and the accessibility of postgraduate students.

In terms of content, the study will examine the extent and patterns of generative AI use, including the purposes for which postgraduate students employ applications such as ChatGPT, Google Gemini, Microsoft Copilot and other relevant generative AI tools. It will also examine students’ information verification behaviour, particularly their practices in checking AI-generated facts, confirming references, consulting original academic publications and comparing information with independent sources.

The study will further examine research integrity in relation to honesty, originality, appropriate acknowledgement, transparency, avoidance of fabricated references and accountability for research content. The respondents will comprise postgraduate students enrolled in the selected universities, including students undertaking postgraduate diploma, master’s and doctoral programmes, where applicable.

The study will not evaluate the technical performance of individual AI applications or compare their internal algorithms. Rather, it will focus on students’ use of generative AI, their verification practices and how these relate to research integrity within the selected university environments.

1.8 Methodology

The study will adopt a descriptive survey research design with a correlational component. The design will be appropriate because the research seeks to obtain information from postgraduate students concerning their use of generative AI, information verification behaviour and research integrity without manipulating the variables.

The population will comprise postgraduate students enrolled in the selected Nigerian universities. A sample will be drawn from the population using an appropriate sampling procedure, taking into consideration the postgraduate student population and the distribution of respondents across the selected institutions.

Data will be collected primarily through a structured questionnaire designed in line with the objectives of the study. The questionnaire will contain sections addressing respondents’ demographic information, extent and patterns of generative AI use, information verification behaviour and research integrity practices. Responses will be measured using an appropriate Likert-type scale. The instrument will be subjected to expert assessment to establish its content and face validity, while a pilot study will be conducted to determine its reliability before the main administration.

The collected data will be analysed using descriptive and inferential statistics. Frequency counts, percentages, means and standard deviations will be used to answer questions relating to the extent and patterns of generative AI use and information verification behaviour. Appropriate inferential statistical techniques, such as Pearson Product Moment Correlation, will be used to examine the relationships among the study variables, subject to the nature and distribution of the collected data. The null hypothesis will be tested at a 0.05 level of significance.

1.9 Operational Definition of Terms

Generative Artificial Intelligence (GenAI): In this study, generative AI refers to digital artificial intelligence systems capable of producing new textual, visual or other forms of content in response to users’ instructions. Examples include ChatGPT, Google Gemini, Microsoft Copilot and Claude.

Generative AI Use: This refers to the extent, frequency and purposes for which postgraduate students employ generative AI applications in their academic and research activities, including idea generation, literature exploration, summarisation, academic writing, language improvement and research assistance.

Information Verification Behaviour: This refers to the actions postgraduate students undertake to establish the accuracy, credibility and relevance of information generated by AI tools. It includes checking original publications, confirming references, consulting scholarly databases, comparing information across sources and examining the evidence supporting AI-generated claims.

Research Integrity: This refers to the commitment of postgraduate students to honesty, transparency, originality, accountability and responsible conduct throughout the research process, including accurate representation of information, appropriate acknowledgement of sources and avoidance of fabrication, falsification and plagiarism.

Postgraduate Students: These are individuals enrolled in postgraduate diploma, master’s or doctoral programmes in the selected Nigerian universities.

AI Hallucination: This refers to instances in which a generative AI system produces information that appears credible but is inaccurate, fabricated, misleading or unsupported by verifiable evidence.

Academic Disclosure: This refers to the transparent acknowledgement of the use of generative AI in academic or research work in accordance with applicable institutional guidelines.

Selected Nigerian Universities: These refer to the Nigerian universities chosen as the institutional locations for collecting data from postgraduate students in this study.

Project – Generative AI Use, Information Verification Behaviour and Research Integrity Among Postgraduate Students: A Study of Selected Nigerian Universities.