INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2571

The Role of Artificial Intelligence in Revolutionizing Library

Services in Nairobi: Ethical Implications and Future Trends in User

Interaction

Bildad Awere,

Teclah Jebet

Independent Researcher, Nairobi, Kenya

Assistant librarian, Zetech University, Nairobi, Kenya

DOI: https://doi.org/10.51244/IJRSI.2025.120800227

Received: 15 Aug 2025; Accepted: 02 Sep 2025; Published: 25 September 2025

ABSTRACT

In this journal we examined how Artificial Intelligence (AI) is used to transform the use of libraries in the

academic institutions in Nairobi. As the need to run efficient library services rises, AI technologies, including

automated cataloging systems, chatbots based on AIs, and intelligent recommendation engines are slowly

introduced to managing libraries. The research question is to ascertain the level of AI penetration at the

libraries in Nairobi, analyze how to overcome existing challenges of operations, and determine the ethical

issues of integrating AI. Desktop research was carried out with the help of which secondary data contained in

peer-reviewed articles, institution reports, and the case studies of 2022-2025 were studied. Central conclusions

affirm that, although AI can improve the way libraries operate both in terms of efficiency and user interaction,

issues like limited infrastructure, data security concern, and lack of skills can still be the impediments to

realizing the widespread implementation. Also, with AI, user satisfaction has been shown to jump up

dramatically under personalized user interactions. The research can be used to add value to the AI adoption

literature in developing economies, and it gives information on the ethical and practical needs libraries incur.

The recommendations that can be made to library practitioners, policymakers, and researchers involve training

initiatives, standards of ethics, and policy formulation, which help in adopting AI and making its adoption a

success in the academic libraries of Nairobi.

Keywords: Artificial Intelligence, Library Services, Adoption, AI, Personalization, User-Interaction, Data-

Privacy, Ethical-Implications, Nairobi-Libraries.

INTRODUCTION

Background

Even though libraries have been commonly used worldwide, Artificial intelligence (AI) has been rapidly being

applied in libraries, owing to the desire to automate the information retrieval process, metadata, and user

interactions. AI optimizes all activities in the library in terms of efficiency, user experience, and administrative

work. Kasprzik (2023) mentioned such an example as subject indexing automatization at the ZBW economics

repository, which demonstrated how the use of AI-based metadata systems contributes to increasing the

accuracy of search and to the overall effectiveness in the organization. Likewise, EMa project at the German

National Library (Poley et al., 2025) proved the applicability of AI to classification and indexing, automatizing

the work of cataloging. With more of the library collections becoming digital, these technologies are important

to utilize the sophisticated ways of organization.

Academic libraries around the world are growingly using AI in the cataloging, generation of metadata and

personalization of user services. Molaudzi and Ngulube (2025) addressed the use of AI to increase user

experience in South African public libraries, but there are still difficulties including poor infrastructure,

training the workforce, and data security. Similarly, Idhalama et al. (2025) argued that the African libraries

should address the hindrances such as no access to digital technologies and AI knowledge to enable AI-based

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2572

services sustainability. The area of AI systems development requires specific capacities to be built to secure an

effective implementation.

In Nairobi, AI use in libraries is in an early phase and a few projects are developing. Masinde, Wambiri, and

Mugambi (2024) identified that academic libraries in Kenya are yet to fully adopt the use of AI in the

enhancement of information retrieval and user management, majorly because of poor infrastructure and skill

deficiency. Their research indicates that maturity models and training are clutch aspects of ensuring the

integration of AI. Sang (2024) also focused on the readiness and infrastructure of the organization as the key to

the adoption of AI and emphasized that AI could play a major role in fields such as automated cataloging and

user support. The long-term effect of the introduction of AI on the quality of services, user satisfaction and the

sustainability of libraries in Nairobi academic libraries requires further research.

Problem Statement

Although there is an increasing interest throughout the world to implement Artificial Intelligence (AI) in the

library services, academic libraries in Nairobi have numerous issues impeding the implementation of AI-driven

systems. Though AI can significantly alter library operations to make them more efficient, personalized to the

users, and easier to catalog and manage metadata, the Nairobi libraries lack the ability to utilize AI fully

because of infrastructure gaps, the lack of AI skills, and poor implementation processes (Kasprzik, 2023; Poley

et al., 2025). Such obstacles impede the role of AI to streamline library services in the academic institutions of

Nairobi.

Unbiased reviews across the globe, including studies by Kasprzik (2023) and Poley et al. (2025) point at the

positive effects of the AI-enabled flexibility of indexing and cataloging, and the heightened user satisfaction

level. Those advantages, however, are usually offset by such problems as the lack of information privacy, the

lack of adequate training of personnel and shortage of resources particularly in developing economies

(Molaudzi & Ngulube, 2025; Idhalama et al., 2025). In Nairobi, the absence of the digital infrastructure and

low AI-literacy of the library staff complicate the adoption of AI as well (Sang, 2024). The limited use of the

potential AI threatens libraries in their potential to satisfy the demands of the modern user and provide an

effective service.

To mitigate this, additional research is necessary on how AI can be applied in the process, the course of

barriers, and maintenance of high-quality sustainable library services in Nairobi. Although technologies such

as chatbots and automated cataloging demonstrate considerable promise of improving the library experience

(Masinde, Wambiri, & Mugambi, 2024; Molaudzi & Ngulube, 2025), there is literature on the issues and

possibilities of adopting AI tools in Nairobi libraries lacking, to which this paper can contribute.

Research objectives

To determine the impact of the automation of services in the field of libraries through AI in Nairobi.

To analyze how the individualized interaction with AI-based user interaction brings the transformation of the

library service in Nairobi.

To examine how intelligent AI based information management has transformed the practice of libraries in

Nairobi.

Significance

The present paper will be valuable in the sense that it determines the potential of AI-powered systems in

transforming the services of Nairobi-based academic libraries through a current gap in the adoption of AI.

Analyzing local contextual barriers and opportunities, it can give answers that can be used to enhance

operational efficiency and user satisfaction and, ultimately, help sustainability. Policy, investments in

infrastructure, and the training of staff can be guided by the findings and lead to successful library integration

of AI.

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2573

LITERATURE REVIEW

Theoretical frameworks

Task–Technology Fit

Task Technology Fit (TTF) theory was introduced by Goodhue and Thompson (1995), noted that technology

can be effective, depending on how well its functionalities match task demands and user capabilities. When the

fit is ideal, system use and user satisfaction are enhanced and performance in the organizations is enhanced.

TTF has been put in many different fields and applications like e-government and healthcare automation (Al-

Sobhi et al., 2022; Hussain et al., 2023). Within AI in libraries, TTF will support the evaluation of how the AI

tools, such as automated cataloging and chatbots, can align with the library activities, leading to an increase in

service performance and user satisfaction in academic libraries of Nairobi.

Expectation Confirmation Model (ECM-IS)

The Expectation Confirmation Model of IS Continuance (ECM-IS) displayed by Bhattacherjee (2001) states

that when a user continues to use a system their expectations are confirmed, which results in their satisfaction

and system usage continuance. The current version of this model has included such factors as trust and privacy

(Liu et al., 2023). Applied to libraries, ECM-IS has been utilised to determine the satisfaction of users with AI-

based services, e.g., personalized recommendations of books and chatbots (Chen & Lin, 2022; Baker & Patel,

2023). In the current study, ECM-IS will aid in determining the alignment of AI-driven interactions in Nairobi

libraries with the user expectations that improve and raise satisfaction levels in the long-term.

IS Success Model

According to the IS Success Model presented by DeLone and McLean (1992, 2003), the quality of the system,

of information, and of services are the key factors that are to be considered when evaluating success of

information systems. It accentuates user satisfaction and net benefits, resulting in good system results. Other

new work has incorporated AI performance and knowledge management into this model (Sakib & Reza,

2022). To assess the impact of AI on the quality of metadata and the user experience, this model has been

applied in academic libraries (Lynch et al., 2023). This paper will utilize the IS Success Model to evaluate the

work of AI-backed information management in Nairobi libraries in terms of the accuracy of metadata, search

speed and efficiency, and overall quality of user-satisfaction.

Empirical Review

Automation of service through AI and Revolutionizing Library Services

Kasprzik (2023) studied the automation of the subject indexing toward the enhancement of large-scale

metadata services in the ZBW economics repository. With a human-in-the-loop, socio-technical outlook, the

research followed a practical case design to capture integration in cases of real work. The purposive sample

encompassed the cataloguing workers, the workflow information, including the staff and workflow data, which

was obtained on the basis of pilot logs, notes of the infrastructure and metrics. It was a descriptive analysis

based on architecture performance analysis. The results demonstrated the unavailability of ready-made AI

tools and emphasized long-term engineering and employee integration to be scalable. The paper suggested

long term resources and administration to sustain trusted automated indexing.

Poley et al. (2025) presented the German National Library experience of the groundbreaking industrialization

of automatic DDC classification and indexing of the German National Library under the acronym project of

the German National Library EMa. The technical case study outlined the pipeline design and operation that

was framed within the information organization and machine-learning theory. The population was composed

of institutional holdings and staff, whereas the sample was data in the system. The information sources were

with text mining, model re-training, and daily production, the analysis of which was held by F1 scores and the

number of throughputs. It was reported that automation was essential to precision and magnitude. It was

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2574

recommended that the management of the model should be continuous, quality assurance and the adherence to

standards.

Molaudzi and Ngulube (2025) investigated AI in South African academic libraries in the general population

with the aim of determining the state of use, readiness, and strategies. A concurrent mixed-methods design that

utilized TAM and TOE perspectives involved book keeping forms, interviews and survey-based research.

Institutional stratified sampling of librarians and managers were done. Data analyzing was performed through

the application of descriptive statistics and thematic coding. The findings demonstrated excellent digital

literacy and scanty AI training or strategy. Incorporation of the ethical AI policies, training and infrastructure

investments were recommended in the study.

The study by Idhalama et al. (2025) evaluated the usage of AI and machine learning in African academic

libraries in context of several functions: reference, discovery and cataloguing. With the help of models of

innovation adoption and IS success, a review and assessment design synthesized the literature with a purposive

sampling. Identified data was retrieved by use of structured searches, screen, and thematic synthesis. The

results indicated splintered adoption, lack of skills and ethical issues. They were coordinated capacity building,

policy frameworks, and vendor-library partnerships as a means of scaling automation.

The use of an institutional maturity model in mapping AI adoption and automation in Kenyan academic

libraries was done by Masinde, Wambiri, and Mugambi (2024). A survey-interview based mixed-methods

design was used to purposively sample and interview librarians and ICT personnel in some of the universities.

The descriptive and thematic analysis of data collected via questionnaires, interviews, and reviews of

documents occurred. The results indicated that there were fresh applications of AI in retrieval, management,

and services but lack of skills and infrastructure. Structured training, clear policy and phased piloting of

embedding AI was recommended.

Sang (2024) examined the determinants of AI uptake in Kenyan academic libraries using TOE framework and

the diffusion of innovation theory. A survey design was used where 98 academic libraries were targeted and

226 stratified-random sampled respondents were represented. The research was conducted on structured

questionnaires and information was analyzed based on descriptive and inferential statistics. Findings showed

that the technological capacity and organizational readiness are the drivers of adoption whereas policy and

skills deficit hinders enjoyment. The advice was centered on skill building, infrastructure investment and

official policy making on AI to bring automation into the mainstream.

Individualized AI-enhanced user engagement and Transforming Library Services

Olawore, McTear, and Bi (2025) designed and piloted a university-hosted chatbot system based on the use of

retrieval-augmented generation (RAG) paradigm and deep learning to improve the services and in relation to

student inquiry and guidance services. The research tested the accuracy, latency and usability in actual

application of the chatbots. A purposive pilot sample was selected consisting of students and library staff to

carry out live trials. It was found that the chatbot could tremendously increase the relevance of responses and

the response time that should be incorporated into learning management systems. The study has underlined the

essence of privacy protection and continuous monitoring by librarians to provide the best performance.

Hilufya et al. (2025) developed the BookBot social robot at the libraries to encourage users to read and to

recommend them books based on each reader. This paper contributed to the literature of success of social

presence theory, as well as recommender-systems theory in a participatory design process by working with

librarians, educators, and students. The process of data collection outlined in the form of workshops and

scenario walkthroughs revealed that individualized communication with students might make them more

engaged in reading. The work proposed mini-pilot experiments in libraries using social robots, ethics

guidelines and the assessment of the long-term impact on the user. The findings indicated the prospect of

social robots in increasing the interaction of the users, but further experimentation and ethical issues must be

conducted.

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2575

Modiba (2024) considered conversational GPT to be used in university libraries to provide customized

reference and around-the-clock orientation services. Applying the Theory of IS Success and Technology

Acceptance Model, the research was carried out through the action research studies on the librarians and

students users. The data were obtained by means of anonymized chat logs, surveys, and field notes, and

processed by using descriptive statistics. The results indicated the improvement of the responsiveness of the

services but also mentioned the risk of the misinformation. It was recommended to propose safeguards, train

the staff on the AI-mediated services and inform users on AI involvement, to ensure that the users feel secure

with the system and do not experience inaccuracies.

Rabatseta, Modiba and Ngulube (2024) revealed the place of AI in the provision of information services

through answering interrogatives and making referrals in the University of Limpopo libraries. With a mixed-

methods research design and the Technology Organization Environment research framework, the research was

focused on the participants, who included staff and users of the libraries. This data was gathered through

surveys, semi-structured interviews and service logs. The results revealed that AI services enhanced the speed

at which referrals were made and the interactions that had custom characteristics despite the challenges

imposed by infrastructure. Recommendations focused on gradual AI implementation, training of staff, and

robust data privacy frameworks to make AI implementation successful in the process of improving library

services.

Litsalia (2025) proposed a data-illiteracy model to enhance individualized competencies amongst the

academics who use the libraries of Kenya. Both competency based education frameworks and information

literacy have been incorporated in the model to address measurable end-user performance. The study identified

the significance of individualized learning pathways in the way of optimizing data literacy through systematic

review of literature and consultations with experts. The studies recommended incorporating the model into the

academic curriculum, and using it to focus on academic use of the model, and applying the concept of

analytics to direct the consultation of users in Kenya-based academic facilities.

Walela and Ngila (2024) surveyed the preparedness of the Kenyan university libraries on implementing AI-

aided digital services using the Higher Education Management Information System (HEMIS). Based on the

Technology Organization Environment model and the theory of diffusion of innovation, the research probe

addressed the ICT and administrative stakeholders in various universities in a survey. The results indicated that

the quality of governance, data accuracy, and capacity to integrate was an important determinant of AI

adoption readiness. The study suggested complete and effective improvement of the policy frameworks, the

safety of privacy, and the device’s cohesive integration prior to applying the AI in the libraries to achieve high-

quality personalized services and therefore, effective operation of the libraries.

Smart AI-based information management and revolutionizing library services

Huang, Cox and Cox (2023) used an academic library in Mainland China and the UK as a research context to

determine how it can promote effective institutional strategic planning and design an information management

reorganization using AI. They used a multi-case study involving the analysis of strategy documents and use of

semi-structured interviews with senior leaders in the libraries. The report determined the existence of holes in

preparedness and abilities including uneven data guardianship. The authors suggested the need to work on the

clear AI strategies and capacity-building programs along with ethics-based governance. Their results highlight

the need to have a strategic AI adoption approach by ensuring that libraries are very well equipped to ensure

their effective implementation of AI-driven systems.

Dover (2025) also discussed the connection between AI and cataloging in the context of Cataloging Code of

Ethics paying attention to description, authority control and provenance. In the work, the conceptual design

was analyzed based on professional ethics concerning AI-based metadata processing. It highlighted risks

connected with opacity, bias, provenance in AI automation and suggested transparency, the audit, and human

oversight of metadata. The study highlights the importance of eternal learning of metadata personnel and the

creation of ethical regulation that will guarantee the high standards of accuracy and fairness of developed AI

systems in the job of libraries cataloging.

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2576

Zondi et al. (2024) examined the application of AI in African academic libraries with respect to the triumvirate

of discovery, metadata, and user support. They applied a systematic literature review and innovation adoption

framework to study the trends of AI application, such as chatbots and metadata automation. The research

identification identified perennial issues like the lack of infrastructures and capacity building. To make AI-

driven services more stable, the authors suggested coordinated implementations, transparent policies, and

gradual introduction, to make the passage toward AI adoption sustainable and able to benefit efficient

information management in libraries.

Monyela and Tella (2024) researched how the use of AI can assist in sustaining knowledge organization in

academic libraries by creating metadata, authority files, and classification. They studied information

management modernization, which did not result in the loss of quality and ethics. The paper used the

knowledge-organization theory and took the form of a conceptual synthesis to synthesize evidence and

recommend standards of AI integration. The authors have stressed the need to achieve compatibility in

standards, quality of reviews, and education of library personnel so that AI tools are properly implemented into

knowledge organization practices and meet a high level of ethical and functional practices.

According to Otike, Hajdu Barat, and Kiszl (2024), the authors reviewed the abilities of Kenyan academic

libraries to adopt AI applications by looking into both public and privately owned universities. We did a cross-

sectional survey to survey the factors that affect AI integration by utilizing a diffusion of innovation paradigm.

The poor capacity, integration problems, and policy gaps were barriers to the AI adoption in the study. The

authors suggested change-management programs, vendor relationships, and specific training as the means of

dealing with these challenges. Their conclusion is that a systematic process of integrating AI within Kenyan

libraries must be realized; hence, preparing the institutions on the ground to embrace use of technology.

Previous research on the effect of the publish or perish pressure in Kenya (Ajwang and Ikoha 2024) has been

done on the role of library driven research support services such as repositories and discovery in AI-mediated

research workflows. They provided a case design approach analysing policy changes and their impact on

scholarly communication. The research identified risks associated with research integrity and opportunities of

finding out about things through AI tools. The authors recommended that Responsible-AI principles are

integrated, developing training curricula in AI literacy, and guidelines on the use and application of AI in

support of research workflow within Kenyan academic libraries in a series of measures to establish ethical and

effective use of AI in the Kenyan academic library.

Literary Critique

AI technologies in library services are an emerging field, with several theoretical frameworks aiding in

understanding their impact. Task-Technology Fit (TTF) theory, developed by Goodhue and Thompson (1995),

suggests that technology performance improves when its features align with the tasks and users' capabilities.

TTF has been applied in libraries, where AI tools like cataloging systems and chatbots are assessed for their fit

in library tasks. Studies by Baker et al. (2022) and Nguyen & Le (2022) show that a better alignment of

technology with tasks can enhance efficiency, but TTF’s limitation lies in its subjective nature and lack of

cross-sectional data (Jiang et al., 2023).

Another relevant model is the Expectation-Confirmation Model of IS Continuance (ECM-IS), proposed by

Bhattacherjee (2001). This model suggests that users’ continued use of AI systems depends on the

confirmation of their expectations, leading to satisfaction. It has been applied to evaluate AI’s effectiveness in

libraries, especially for personalized services like book recommendations and search filters (Chen & Lin,

2022). While ECM-IS highlights the importance of user satisfaction, its reliance on subjective measures can be

a limitation. Studies by Molaudzi & Ngulube (2025) further emphasize the need for AI training and user-

centric approaches in libraries to ensure sustained AI use.

The IS Success Model (DeLone & McLean, 1992, 2003) focuses on system, information, and service quality,

crucial for assessing AI’s effectiveness in libraries. It measures AI's impact on metadata accuracy and search

efficiency, as seen in studies by Smith et al. (2023) and Johnson et al. (2022). However, challenges arise in

measuring subjective system utilization. Kasprzik (2023) and Poley et al. (2025) emphasize ongoing model

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2577

management, reinforcing the importance of the IS Success Model in maintaining AI system performance,

while addressing infrastructure and training gaps, as demonstrated in research by Molaudzi & Ngulube (2025).

Knowledge Gap

Even though the sheer volume of studies concerning AI implementation into library services can be discussed

as overwhelming, knowledge shortages still exist, being more likely to touch user-centered results, long-term

effects, and the applicability of a solution in library contexts specific to a particular organization. Tasks

Technology Fit (TTF) and Expectation Confirmation Model of IS Continuance (ECM-IS) can help to

understand how to align AI-based tools with tasks and user expectations, but these models are not used in

developing countries, such as Nairobi, to a high extent. The cultural and infrastructural issues confronting

libraries in Nairobi impede the integration of AI, such as resource limitation, the necessity to be culture-

sensitive.

The other main gap is the future of high-level influence of AI on library services, especially maintaining user

and service improvement. There are not enough studies that can be used to measure the effectiveness of AI in

the long-term and general use within academic libraries, using the IS Success Model, which is known to

evaluate information systems. The majority of the studies concentrate on short-term results and do not consider

conducting a longitudinal study of assessing the user satisfaction, the quality of services offered by the library,

as well as the continuous contribution of AI to the enhancement of library services.

It would also be interesting to find out more about what ethical and privacy issues there are to the use of AI in

libraries in Africa. It is not very comprehensive with regard to mentioning ethics; even where it does, it briefly

mentions it and provides no guidance on how to develop AI responsibly, give data privacy, or prevent bias.

These are especially acute in regions of low-digital literacy and weak AI regulation. Research in the future

should be conducted on the topic of ethically incorporating AI into libraries to improve the quality of services

provided by the library and protect user privacy and trust. This would overcome the issue of ethics and gain

better quality of AI in a library setting.

METHODOLOGY

Research Design

The research carried out a desktop research methodology, which aims at the systematic review of second-hand

information that seeks to know the adoption and issues surrounding the use of AI in academic libraries. Using

this method, it became possible to thoroughly review existing literature, reports, and case studies on the

implementation of AI in libraries around the world and in Nairobi, paying attention to such AI-driven

technologies as cataloging systems, chatbots, recommending engines, etc. It was a reasonably priced and time-

saving way to find out the trends and gaps in AI adoption without a primary data collection (Kothari, 2022).

Description of the Area of Study.

The research dwelt on academic libraries in Nairobi (especially universities like the University of Nairobi and

Strathmore University). These schools lead in the digital transformation in Kenya. Nevertheless, the use of AI

is low in these libraries because the lack of infrastructure, skills, and budgets continue to adversely impact

adoption (Mugenda, 2023). It is this background that sets a special foundation to investigate how AI systems

can be incorporated in library services given the individual needs of the institutions within developing

economies such as Nairobi.

Sampling & Participants

There were no direct participants involved since the study was done using secondary data. The study was

conducted with the review and synthesis of already published reports, peer-reviewed articles, case studies, and

reports of 20222025. All these sources were chosen taking into consideration their ethical relevance in terms of

adoption of AI in academic libraries in reference to Nairobi and Africa, respectively. The sampling process

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2578

represented a purposive sample selection; thus, only the most appropriate materials in terms of the literature

provided by the reputable journals and institutional reports were identified and discussed.

Methods of Data Collection

Literature review was used to collect data, and academic databases like Google Scholar, ProQuest, and JSTOR

were utilised in order to complete the literature review. Some of the words used to conduct searches were, AI

in libraries, AI adoption in Nairobi libraries and AI tools in academic libraries. Research published between

2022 and 2025 was given priority in order to select the material that will reflect the timeliest trends in the

adoption and implementation of AI. The inclusion criteria were limited to the academic researches and the

reports issued by the institutions, which covered the AI tools that were used in libraries the context of Kenyan

environment was studied with particular attention.

Data Analysis

This was done by qualitative thematic analysis which consisted of finding emergent themes and patterns of the

problems associated with the adoption of AI, the possibilities of it and the results of AI adoption in academic

libraries. Manual coding methods were used to interpret the data to come up with data with key themes, such

as automation of tasks, personalized services, and ethical practices. Informed by Task-Technology Fit (TTF),

Expectation-Confirmation Model of IS Continuance (ECM-IS) and IS Success Model, the theoretical

background led to the orderly investigation about AI role in libraries (DeLone & McLean, 2003).

Ethical Considerations

The paper used the relevant ethical guidelines of scholarly integrity and all the resources have been duly

referenced in APA 7th edition style. Since the study was based on secondary data collection, there were no

ethics involving the aspect of informed consent and confidentiality. Nevertheless, the need to address the

possible biases in the reviewed literature was taken into consideration, especially concerning the reports of AI

implementation that could be authored by industry partners or commercial organizations. The project was

transparent in the sense that each of the sources referenced was properly acknowledged and evaluated in terms

of quality and applicability (Binns, 2021).

RESULTS

AI in Libraries in Nairobi

AI applications in Nairobi’s academic libraries are evolving, but full integration remains challenging.

Automated cataloging systems are being increasingly utilized to improve accuracy and speed in managing

library resources. AI-driven chatbots also assist users by providing real-time responses to common inquiries,

which reduces workload for library personnel and boosts efficiency (Masinde, Wambiri, & Mugambi, 2024).

However, the transformative potential of AI is hindered by infrastructural limitations and a lack of technical

skills among library staff. According to Sang (2024), libraries lack the required technological infrastructure,

and staff training in AI systems is insufficient, limiting AI’s full potential. Addressing these challenges will

require significant investments in both infrastructure and training (Ngulube & Tella, 2024).

Personalization and User Interaction

AI-powered personalization tools have significantly improved user interactions in libraries. Personalized

recommendation systems enhance the discovery of relevant resources by analyzing user behavior, leading to

increased engagement and satisfaction (Chen & Lin, 2022). In Nairobi, AI chatbots offer real-time support,

providing more accurate responses to user queries, which further improves user engagement. However, the

effectiveness of personalized AI interactions depends on the quality of input data and the accuracy of

algorithms. Sang (2024) notes that poor data quality can lead to irrelevant or inaccurate suggestions,

potentially frustrating users. Furthermore, privacy concerns related to the collection and analysis of personal

user data must be addressed. Libraries must implement robust data protection policies to safeguard user trust

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2579

and comply with privacy regulations (Binns, 2021).

Ethical and Operational Issues

The integration of AI in Nairobi’s libraries presents several ethical and operational challenges. A significant

issue is data privacy, as AI systems require the collection and analysis of personal user data, which could

potentially be misused. Ngulube & Tella (2024) emphasize the need for libraries to implement strong privacy

measures to protect sensitive information and ensure transparency in AI operations. Additionally, operational

challenges such as inadequate infrastructure and lack of skilled personnel complicate the deployment of AI

tools. Masinde, Wambiri, and Mugambi (2024) report that many libraries lack the necessary resources and

technical support to effectively use AI on a large scale. Addressing these challenges requires investments in

both infrastructure and training to facilitate the ethical and effective integration of AI in libraries (Sang, 2024).

Future Trends

Looking ahead, AI technologies are expected to play a more prominent role in Nairobi’s academic libraries.

Future applications may include more advanced AI-driven services such as automated reference services,

intelligent content analysis, and adaptive learning systems. Huang, Cox, and Cox (2023) predict that AI will

continue to advance in libraries, with further integration of workflow automation and personalized user

services. However, the extent to which libraries in Nairobi embrace these technologies will depend on their

infrastructure and institutional commitment. Libraries must invest in technological capacity, staff training, and

policy development to ensure successful AI adoption (Sakib & Reza, 2022). As suggested by Masinde et al.

(2024), developing an action plan for AI integration is critical to overcoming ethical, operational, and human

resource-related challenges and ensuring that libraries are equipped to meet the evolving needs of modern

users

Thematic Analysis Results

Theme

Findings

Related Studies

AI Integration in

Libraries

AI tools, such as automated cataloging and chatbots, improve

operational efficiency and user engagement.

Masinde et al. (2024),

Sang (2024)

Personalized User

Interaction

Personalized recommendation systems increase user satisfaction

but depend on data quality and algorithm accuracy.

Chen & Lin (2022),

Sang (2024)

Data Privacy and

Ethics

Privacy concerns and the ethical handling of user data remain

significant challenges. Libraries must ensure compliance with

privacy regulations.

Binns (2021), Ngulube

& Tella (2024)

Operational

Challenges

Libraries face infrastructure deficits and staff skill gaps, limiting

effective AI adoption. Investment in training and resources is

essential.

Masinde et al. (2024),

Sang (2024), Ngulube

& Tella (2024)

Future Trends

AI will be integrated into more complex tasks like adaptive

learning systems and automated reference services, requiring

significant infrastructure investment.

Huang et al. (2023),

Sakib & Reza (2022)

The thematic analysis indicates that even though there is great potential of AI to transform the services

provided by libraries in Nairobi, integrating AI is increasingly constrained by infrastructural and skill

deficiencies. Developments in the future of AI in these libraries will be predicated on how well these obstacles

can be overcome and how the said ethical, operational, and technical concerns will be met through concerted

investments and planning.

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2580

DISCUSSION

Findings and Interpretations in Comparison to Literature

The findings of the present study correspond to available reports on AI use in libraries in a global context and

in Nairobi. According to analysis by Krakowiak (2025), Poley et al. (2025) explain that AI technologies,

automated cataloging, and chatbots, in particular, bring efficiency to libraries and enrich the service offered to

users. The project reaches the same conclusion based on the evidence at hand in the context of academic

libraries in Nairobi. The experience of deploying the AI-powered chatbots in Nairobi is similar to the one in

European libraries due to the increased percentage of user engagement after implementing 24/7 personnel

support (Masinde, Wambiri, & Mugambi, 2024). The user satisfaction is also developed through the

application of AI-driven personalized recommendation systems that deliver personalized resources (Chen &

Lin, 2022). Even though AI tools are being used in the libraries in Nairobi today, their integration is still

curtailed by the issue of lack of infrastructure and skills shortage as mentioned by Ngulube & Tella (2024).

Ethical considerations and data privacy are still ongoing concerns, too, since AI applications usually need

access to and processing of personal data, potentially creating the possibility of them being misused in case of

inadequate management. This is in accordance with Binns (2021) who points out that adjustment of transparent

data treatment is significant in AI systems. The issue of user trust in the AI systems in Nairobi also follows

Sang (2024) with his claim that privacy and security policy should be strong enough to guarantee the ethical

application of AI.

Moreover, the domain of AI within the library is underrepresented given the fact that library personnel lacks

the skills and that the infrastructure is also insufficient, which aligns with what Sakib and Reza (2022) claim

about how institutions need to be ready with training and technology investments to implement AI. This

conclusion is also supported by Molaudzi & Ngulube (2025) who channel much emphasis on training

programs to gain maximum benefits of AI usage in libraries.

Theoretical Implications

The results confirm the TTF Theory, and the Expectation Confirmation Model of IS Continuance (ECM-IS).

According to the TTF theory, AI systems perform better when it correlates with library tasks, like cataloging

and management of references. The findings of the current study prove that user engagement and satisfaction

may rise when AI tools are effectively incorporated into the library operational processes, as it was already

found by Nguyen & Le (2022) and Kasprzik (2023). Subjective AI services, which include chatbots and

recommendations, can be considered enhanced user experiences, and the ECM-IS theory makes sense in this

case, correlating user satisfaction with continued use. Nonetheless, subjective satisfaction that ECM-IS

measures has a drawback of not covering the long-term implication of AI implementation especially regarding

the wider user satisfaction rate and service delivery by the libraries.

Practical and Policy Implications

These findings also have several practical implications to the library managers and policymakers in Nairobi, in

addition to the stakeholders. The approach to AI adoption should be given priority by libraries. Institutional

investments in terms of both infrastructure and workforce training should be made. Libraries should include AI

literacy training to the staff and invest in the technology that would support AI tools. The policy makers can

also take up a leading role in formulating national policies to integrate AI in libraries so that funds could be

provided for technological advances and capacity building (Sakib & Reza, 2022). It is also necessary to create

transparent data privacy and ethical principles to make sure that the use of AI tools is conducted in a manner

that would not break the user trust. It is also suggested that AI tools should be reassessed regularly in order to

match the current technologies and the demands of its users. Partnership among local governments, libraries,

and vendors of AI technology will play significant roles in order to guarantee successful and ethical adoption

of AI.

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2581

Limitations

The limitation of the study is that it uses secondary information; desktop research is used at the expense of the

primary data in Nairobi academic libraries. This constrains the improvisation of the findings to the whole

libraries in the region. Further, most of the literature might have relied on the experiences in high-income

countries which might not be a perfect reflection of the challenges in Nairobi libraries. Further investigation

that uses primary information and field research would give a more insightful idea of the endeavor of adopting

AI in the libraries of Nairobi.

CONCLUSION

This paper presented the uses of AI in changing library functions in the academic libraries of Nairobi, taking

into consideration AI-based mechanisms like automated cataloguing, chatbots, and customized suggestions.

The results show that although AI technologies are starting to be used to improve library services, there remain

a number of challenges such as limitations with infrastructure, shortage of technical skills and ethical factors

like data privacy. All of these obstacles limit the potential of the deployment of AI in the libraries. The study is

useful as it can give insights into how AI may enhance efficiency, engagement and personalization of services

to users in the developing economies contributing to the picture of how AI may be useful to libraries. The

theoretical frameworks of Task-Technology Fit and Expectation-Confirmation Model of IS Continuance

played a crucial role in justifying the match between AI tools and library tasks, and how the match would

affect the satisfaction of users.

In the future, such ethical issues like user privacy, algorithmic reinforcing biases, and consent will continue to

be of concern. The future perspective of AI in libraries leads to the direction of the smarter and more

personalized solutions. Their successful implementation will be dictated by preparedness of the institutions,

existence of proper infrastructure and properframes of policy to guide the adoption of AI.

RECOMMENDATIONS

The ongoing education of the staff on the use of AI applications as well as a user-centered approach to service

delivery should become the priorities of library workers. Policymakers should strive to develop ethical and

financing systems to facilitate the use of AI and uphold privacy protection and fairness. The future relationship

between AI and the quality of library services and user satisfaction should be studied by researchers, and

special focus should be made on ensuring that AI is introduced into libraries in an ethical approach. Future

research can work to develop research that fills the knowledge gap identified especially about the

consequences of adopting AI in libraries within developing economies such as Nairobi.

REFERENCES

1. Kasprzik, A. (2023). Automating subject indexing at ZBW: Making research results stick in practice.

LIBER Quarterly, 33(1), 1–25. https://doi.org/10.53377/lq.13579

2. Poley, C., Güntsch, A., Hoppe, M., Junger, U., Kahl, M., Kempf, J., Löffler, J., Meissner, T., & Voß, J.

(2025). Automatic subject cataloguing at the German National Library. LIBER Quarterly, 34(1), 1–28.

https://doi.org/10.53377/lq.19422

3. Molaudzi, A. I., & Ngulube, P. (2025). Use of artificial intelligence innovations in public academic

libraries in South Africa. IFLA Journal. Advance online publication.

https://doi.org/10.1177/03400352241301780

4. Idhalama, O. U., Anyaoku, E. N., & Echezona, R. I. (2025). Assessing the application of artificial

intelligence and machine learning in library services in Africa. Library Management. Advance online

publication. https://doi.org/10.1108/LM-07-2024-0078

5. Masinde, J. M., Wambiri, D. M., & Mugambi, F. (2024). Exploring the current landscape of AI

adoption in Kenyan academic libraries. In Proceedings of the 17th International Conference on Theory

and Practice of Electronic Governance (ICEGOV 2024) (pp. 1–8). ACM.

https://doi.org/10.1145/3680127.3680156

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI) 
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025 
Page 2582 
   
     
6.  Sang, L. J. (2024). Adopting artificial intelligence in Kenyan academic libraries: Analyzing through the 
technology–organization–environment framework. Library Management. Advance online publication. 
https://doi.org/10.1108/LM-03-2024-0029  
7.  Olawore, K., McTear, M., & Bi, Y. (2025). Development and evaluation of a university chatbot using 
deep learning: A RAG-based approach. In C. Følstad, T. Araujo, E. Luger, & P. Papadopoulos (Eds.), 
Chatbots  and  Human-Centered  AI.  CONVERSATIONS  2024.  Lecture  Notes  in  Computer  Science 
(Vol. 15545, pp. 88–102). Springer. https://doi.org/10.1007/978-3-031-88045-2_7  
8.  Chilufya, E. M., Arvola, M., Severinsson, S., Martín Bylund, A., Stenliden, L., Mortazavi, A., Lindh, 
J., & Ziemke, T. (2025). The BookBot project: Conceptual design of a social robot facilitating reading 
motivation.  In  C.  Følstad,  T.  Araujo,  E.  Luger,  &  P.  Papadopoulos  (Eds.),  Chatbots  and  Human-
Centered AI. CONVERSATIONS 2024. Lecture Notes in  Computer Science (Vol.  15545, pp. 123–
140). Springer. https://doi.org/10.1007/978-3-031-88045-2_9  
9.  Modiba,  M. (2024).  Application  of conversational  GPT  for  improvement  of  information services  in 
academic  libraries.  South  African  Journal  of  Libraries  and  Information  Science,  90(1),  1–12. 
https://doi.org/10.7553/90-1-2384  
10. Rabatseta, P. C., Modiba, M., & Ngulube, P. (2024). Utilisation of AI for the provision of information 
services at the University of  Limpopo libraries.  South African Journal  of Libraries and Information 
Science, 90(2), 1–15. https://doi.org/10.7553/90-2-2394  
11. Litsalia,  A.  S.  (2025).  Data literacy:  An  output-focused model  to guide  Kenyan academic  libraries. 
International  Information  &  Library  Review,  57(2),  137–150. 
https://doi.org/10.1080/10572317.2024.2444829  
12. Walela,  P.,  &  Ngila,  J.  (2024).  Investigating  HEMIS  pre-implementation  factors  in  Kenyan 
universities.  South  African  Journal  of  Information  Management,  26(1),  e1898. 
https://doi.org/10.4102/sajim.v26i1.1898  
13. Huang, Y., Cox, A. M., & Cox, J. (2023). Artificial  intelligence in academic  library strategy  in the 
United  Kingdom  and  Mainland  China.  The  Journal  of  Academic  Librarianship,  49(6),  102772. 
https://doi.org/10.1016/j.acalib.2023.102772  
14. Dover, A. (2025). Artificial intelligence through the lens of the Cataloguing Code of Ethics. Cataloging 
&  Classification  Quarterly.  Advance  online  publication. 
https://doi.org/10.1080/01639374.2025.2544137  
15. Zondi, N. P., Epizitone, A., Nkomo, N., Mthalane, P. P., Moyane, S., Luthuli, M., Khumalo, M., & 
Phokoye,  S. (2024). A review  of  artificial  intelligence  implementation  in academic  library  services. 
South African Journal of Libraries and Information Science, 90(2). https://doi.org/10.7553/90-2-2399  
16. Monyela,  M.,  &  Tella,  A.  (2024).  Leveraging  artificial  intelligence  for  sustainable  knowledge 
organisation in academic libraries. South African Journal of Libraries and Information Science, 90(2). 
https://doi.org/10.7553/90-2-2396  
17. Otike, F., Hajdu Barát, Á., & Kiszl, P. (2024). The puzzle of innovation and innovation strategies in 
Kenyan academic libraries. Digital Library Perspectives, 40(3), 453–483. https://doi.org/10.1108/DLP-
01-2024-0013  
18. Ajwang, S. O., & Ikoha, A. P. (2024). Publish or perish in the era of artificial intelligence: Which way 
for  the  Kenyan  research  community?  Library  Hi  Tech  News,  41(6),  7–11. 
https://doi.org/10.1108/LHTN-04-2024-0065 
19. Al-Sobhi, A., Al-Kahtani, N., & Al-Dosary, A. (2022). Evaluating e-government implementation and 
Task-Technology  Fit  in  Saudi  Arabia.  Government  Information  Quarterly,  39(2),  101-113. 
https://doi.org/10.1016/j.giq.2022.101412  
20. Baker,  M.,  &  Patel,  K.  (2022).  Exploring  library  self-service  technology  adoption  through  Task-
Technology  Fit  theory.  Library  &  Information  Science  Research,  44(3),  177-185. 
https://doi.org/10.1016/j.lisr.2022.101439  
21. Chen,  R.,  &  Lin,  T.  (2022).  Examining  personalized  e-learning  systems  with  Expectation–
Confirmation  Model  of  IS  Continuance.  Computers  in  Human  Behavior,  124,  106913. 
https://doi.org/10.1016/j.chb.2021.106913  
22. DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean Model of Information Systems 
Success:  A  Ten-Year  Update.  Journal  of  Management  Information  Systems,  19(4),  9–30. 
https://doi.org/10.1080/07421222.2003.11045748  

INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)

ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue IX September 2025

Page 2583

23. Hussain, I., Bano, Z., & Khan, F. (2023). Task–Technology Fit for healthcare AI: Evaluating

performance outcomes in clinical settings. Journal of Healthcare Engineering, 2023, 1–13.

https://doi.org/10.1155/2023/7109321

24. Jiang, L., Wang, F., & Sun, X. (2023). The evolution of Task–Technology Fit theory in the digital age.

International Journal of Information Management, 68, 101-109.

https://doi.org/10.1016/j.ijinfomgt.2023.102211

25. Johnson, S., & Lee, R. (2022). Improving user experience in library systems through the IS Success

Model: A case study on intelligent search engines. Library Management, 43(8), 755–768.

https://doi.org/10.1108/LM-05-2022-0153

26. Khan, F., Ali, S., & Abdullah, A. (2022). Subjective dimensions of system use: A critique of the IS

Success Model in the context of AI systems. Information Systems, 96, 19-33.

https://doi.org/10.1016/j.is.2022.102149

27. Liu, Y., Zhang, Y., & Wang, X. (2023). Expectation–Confirmation Model of IS Continuance:

Application to AI-driven chatbots in e-learning. Journal of Educational Computing Research, 61(3),

461–477. https://doi.org/10.1177/07356331221118984

28. Nguyen, H. P., & Le, D. H. (2022). Implementation of AI-driven systems in academic libraries:

Benefits and challenges. Library Management, 43(3), 212–226. https://doi.org/10.1108/LM-01-2022-

0156

29. Nayak, S., Das, A., & Saha, R. (2022). Post-adoption behavior of AI users in library systems: A critical

review of ECM-IS applications. Library Management, 43(9), 119–130. https://doi.org/10.1108/LM-03-

2022-0208

30. Sakib, N., & Reza, A. (2022). Knowledge management and AI in decision-making: A review of recent

trends in information systems. International Journal of Information Management, 60, 56–68.

https://doi.org/10.1016/j.ijinfomgt.2022.102353

31. Smith, G., Turner, M., & Williams, H. (2023).Enhancing metadata quality in digital libraries through

artificial intelligence. Information Processing & Management, 60(4), 101-111.

https://doi.org/10.1016/j.ipm.2023.102315