Faculty of Commerce, Law and Management (ETDs)
Permanent URI for this communityhttps://hdl.handle.net/10539/37778
Browse
13 results
Search Results
Item Drivers of the use of AI-powered tools in academic research: A study of university students in South Africa(University of the Witwatersrand, Johannesburg, 2024) Ngomane, Danisile Priscilla; Dorson, Thomas AnningThe integration of Artificial Intelligence (AI) across various sectors, including education, has garnered significant attention due to its potential to enhance productivity, accuracy, and innovation. There has been a notable lack of research focusing on students' adoption of AI tools in research settings, within the context of South Africa. The primary objective of this study was to provide insights to inform strategies aimed at facilitating the effective implementation of AI-powered technologies within the academic research landscape. The study employed a theoretical framework grounded in the Technology Acceptance Model (TAM), which provides a foundation for understanding individuals' acceptance or rejection of technology. The study aimed to investigate the drivers influencing university students' utilisation of AI tools in educational research settings. Adopting a positivist research philosophy, the study utilised a quantitative research design as the primary inquiry strategy. This involved the administration of a research survey to collect and analyse numerical data, enabling the testing of hypotheses, identification of patterns, and quantification of relationships between variables. The study employed a cross-sectional time horizon and utilised convenience sampling to administer a survey questionnaire to 271 final-year undergraduate and postgraduate students at Witwatersrand University. Analysis of the collected data using descriptive and inferential statistics revealed several key findings. A positive attitude towards utilising AI tools was found to be driven by perceived usefulness, perceived ease of use, prior knowledge, and awareness about the tools. This relationship was moderated by factors such as trust and self-efficacy. The AI tools were primarily utilised for tasks such as searching and summarising articles, writing, editing grammar, and making unbiased judgments about articles. Importantly, students were influenced by the perceived benefits derived from using specific AI tools, and the ease of use associated with these tools. This research study findings suggest that promoting awareness of AI tools among university students has the potential to enhance research outcomes, improve scholarly efficiency and increase overall effectiveness. As the demand for AI technologies continues to grow, integrating them into higher education settings holds promise for unlocking substantial benefits and transforming the academic research landscape.Item Examining the influence of digital payment adoption on the digital divide in Kagiso township, South Africa(University of the Witwatersrand, Johannesburg, 2024) Mokgwatsane, Malengolo; Godspower-Akpomiemie, EuphemiaDigital payment systems have become increasingly popular and a vehicle for financial inclusion and access to financial services. However, as technological advancements progress, there remains a population segment that is left behind in terms of adoption and access. Digital payment systems are generally not adopted by township residents in South Africa. This is due to the uneven telecommunications infrastructure in South Africa and its limited availability in the townships. This study investigates the influence of tap-to-pay digital payments on the existing digital divide in Kagiso township and the factors that impact adoption and usage of this technology. This study adopted a qualitative research methodology that involved conducting face-to-face interviews with 12 respondents from Kagiso township. The respondents’ feedback was consolidated and analysed to uncover key findings. The findings revealed the presence of a digital divide in access and availability within Kagiso township, characterised by challenges stemming from poor infrastructure and limited access to technology, worsened by an uneven distribution of ICT. Furthermore, this study revealed lack of awareness and understanding of tap-to-pay, safety and security, unemployment and socioeconomic challenges as significant barriers to the adoption and usage of tap-to-pay payment systems in the township. Therefore, it is advisable for the South African government to increase investment in infrastructural upgrades to improve the accessibility and availability of digital technologies in the townships while also collaborating with the private sector to support initiatives aimed at addressing the digital divide in the townships. The collaboration with the private sector, particularly the telecommunications companies is crucial in driving these efforts. By leveraging market opportunities, telecommunication companies can implement strategies that accelerate the adoption of new technologies. One such strategy is investment in fiber optic networks to provide high-speed internet access in townships.Item Investigating the influence of chatbots on customer experience and frustrations in self-help functions in South Africa(University of the Witwatersrand, Johannesburg, 2023) Raphela, Lesego Jerminah Mmakgopa; Anning, Thomas DorsonOrganisations must provide responsive, efficient, and 24/7 customer service in the ever-evolving digital era. Chatbots have emerged as the preferred Artificial Intelligence (AI) technology, offering self-help functions to users in need of various digital services. This study addressed the scarcity of empirical studies on the use of chatbots as self-service agents in South African companies, particularly in exploring their contribution to customer experiences, both positive and negative. The study extended the Technology Acceptance Model (TAM) to investigate the antecedents of customer chatbot engagement and their subsequent impact on customer satisfaction and frustration within the context of self-help functions. Utilising a quantitative approach, data was collected through online survey questionnaires from a sample of 258 participants who had interacted with chatbots. Multiple linear regression analysis was employed for data analysis. Results revealed that perceived ease of use, performance expectancy, compatibility, and social influence positively influenced customer chatbot engagement. Additionally, customer engagement with chatbots had a significant positive correlation with satisfaction and a negative correlation with frustration. These results suggested that enhancing user-friendly interfaces, ensuring optimal performance, aligning with user preferences, and leveraging social influence could foster increased engagement. This study stressed the significance of understanding and optimising customer chatbot engagement for a positive user experience.Item A feminist ontology to data commercialisation: Evaluating women's access to information and privacy within the medico-legal sphere in South Africa(University of the Witwatersrand, Johannesburg, 2024) Neto, Ângela Pacheco; Swemmer, SheenaWith the dawn of the Fourth Industrial Revolution, rapid exchanges of data have intensified. Technologies like biometric monitoring, female-oriented technologies, and artificial intelligence bring with them a host of legal issues related to consent, access, privacy, and liability. Vulnerable populations or groups must be given particular attention as standard data practices serve to reinforce existing inequalities. For this reason, female-directed and female- generated health data is specifically considered herein. By employing a data feminism lens, it becomes apparent that the current South African regulatory framework has been legislatively misapproached with regards to the medico-legal sphere in South Africa. The methodology herein draws on critical review methods, thematic analysis, and legal discourse analysis, ultimately utilising the general principles of research inherent in the socio-legal sciences. A responsive and flexible health data law that incorporates intersectional narratives is advanced. This holistic response must account for the two-faced coin of female access to information and privacy in order to address historical structures of power inequityItem Artificial intelligence and automated decision making under the GDPR and the POPIA(University of the Witwatersrand, Johannesburg, 2024) Goldman, Gavin David; Zitzke, E.This analysis considers the concepts of AI and machine learning and examines their reliance on the processing of personal data and the challenges this poses from a data- privacy and human-rights perspective, particularly in relation to profiling. It evaluates the effectiveness of the General Data Protection Regulation (GDPR) and the Promotion of Personal Information Act 4 of 2013 (POPIA) in regulating Automated Decision Making (ADM) and considers the limitations of the right to an explanation under these provisions. The analysis proposes that the current framework of the GDPR and POPIA does not clearly address the issue of explainability and that the focus should shift to providing a data subject with a counterfactual to give practical effect to this right which would better serve data subjectsItem A practical review of the effectiveness of artificial intelligence in the automated review of legal contracts(University of the Witwatersrand, Johannesburg, 2023) Fouché, Jacques GerritThis research report compares the performance of AI-driven automated contract review platforms with that of human contract reviewers, aiming to see improved effectiveness for the elements of time, cost and quality. An empirical study is done by comparing the specific performance of an automated contract review platform provider, Lawgeex, to the human contract reviewers of a business entity, Endress+Hauser. The results of the effectiveness assessment are reported on through dashboard data and questionnaires to the users of the platform. Recommendations are made both in general and specific to the two entities of the empirical studyItem Prospects for artificial intelligence to manage load-shedding in South Africa(University of the Witwatersrand, Johannesburg, 2022) Shakoane, Nomea Lerato; Lee, GregoryEskom, a state-owned utility in South Africa, is currently facing significant challenges and experiencing severe power shortages. While there is a growing expectation of adopting renewable energy in the future, a sudden and complete transition is unlikely. Legacy power systems, characterized by poor performance, breakdowns, and unpredictability, have received limited attention in AI research. This raises the question: What actions should be taken to quickly address maintenance issues in older power plants and increase generation capacity in the short term? The objective of this study is to explore AI solutions in the electrical sector and assess the feasibility and cost-effectiveness of integrating AI into Eskom's power system. The findings of this study will provide Eskom and the South African government with valuable insights to make informed decisions regarding the incorporation of artificial intelligence. These AI solutions can include detecting power and cable theft, optimizing energy usage and distribution, and implementing predictive analytics for demand planning and power production optimization. To gather data, a survey questionnaire was distributed to participants primarily located in South Africa, following a snowball selection process. The survey collected responses from a minimum of 50 participants and covered various aspects, such as load shedding at Eskom, artificial intelligence, data-AI enablers, and AI prospects. The study revealed that inadequate maintenance within the power generation division was responsible for load shedding. As a result, the implementation of AI solutions such as predictive maintenance, fault detection, and power demand monitoring systems emerged as crucial priorities for Eskom. However, it is important to note that implementing AI requires substantial capital investment. Considering Eskom's current financial situation and South Africa's mounting debt, it is challenging for Eskom to secure the necessary funds without seeking support vi from the South African government or major corporations like the IMF or World BankItem Exploration of the adoption of Artificial Intelligence within traditional recruitment and selection processes among South African Universities(University of the Witwatersrand, Johannesburg, 2023) Murerwa, GodwinIndustries across the world are investing in and embracing artificial intelligence in their recruitment and selection strategies. Artificial Intelligence (AI) in talent acquisition, fuelled through technological innovation has disrupted traditional recruitment and selection processes. Within South Africa, adoption of AI has been reported to be slow with most organisations still at a pilot stage. An organisation competitive advantage is generated from the people it employs. Due to increased globalisation and interlinked economies, companies rely on their highly talented employees for sustainable competitiveness. It is critical in a world where there is scarcity of skills, talent and resources to use an intelligent system in staffing. Traditional techniques of managing entities are being contested globally. Traditionally, recruitment and selection was a laboriously administrative, mundane and manual process susceptible to prejudices and biases of the practitioner. This research explored the factors influencing Artificial Intelligence adoption in the traditional recruitment and selection practices among South African universities. The main objectives of the study were to establish factors influencing the adoption of Artificial Intelligence in recruitment and selection processes among South African universities and consequently recommend feasible strategies of adopting such technology within the talent acquisition practices of the universities. The study utilised a quantitative research approach through a survey design. For data collection, the research instrument was an online self-completion questionnaire. The study employed non-probability sampling, purposive sampling in particular. This enabled the research to focus on the characteristics and attributes consistent with the study objectives. A total of ninety-five participants from nine South African universities completed the online questionnaire. The findings demonstrated the potential benefits of AI adoption to the hiring processes of South African universities. The benefits of AI adoption seem to be mostly in perceived accuracy and speed in which it curtails tedious tasks in recruitment and selection. There are however significant challenges that should be addressed. Issues to do with ethicsItem Factors Influencing the Adoption of Artificial Intelligence Technologies in the South African Construction Industry(University of the Witwatersrand, Johannesburg, 2023) Mgolombane, Pumza Portia; Matshabaphala, ManamelaThe study aimed to evaluate factors influencing the adoption of artificial intelligence technologies within the South African Construction Industry. The focus was to employ AI technology adoption strategies to enhance the construction industry’s’ performance in ensuring effectiveness and efficiency on productivity through organisational development. Since the construction industry is notorious for its resistance to integrating new technology; for example, it is more likely to continue working manually than to apply digitalisation. As a result, the slow adoption of technology is likely to have an impact on the effectiveness and efficiency of the construction industry building practices. Therefore the studies objectives were to assess, the adoption of AI technology strategies and its impact on effectiveness and efficiency within the construction industry. Moreover, opportunities and constraints of AI adoption by the construction industry to help improve its productivity. Also, the organisational development phenomenon within the construction industry and its effect on AI Technology adoption. This study employed a case study research method to obtain a comprehensive understanding and empirical questions were established in the literature based on the study’s’ aims and objectives. The quantitative data was administered through the Qualtrics tool whereby a survey link was obtained and then emailed to the participants. The data was analysed using Statistical Package for the Social Sciences tool (SPSS) version 27. 51 Respondents participated in the survey however, 9 were excluded because of incomplete surveys. According to the results of the data analysis, it was concluded that the employment of AI technology techniques may have an impact on the effectiveness and efficiency of the building sector, making this objective relevant to this study. Moreover, AI adoption by the construction industry may increase productivity by understanding the opportunities and limitations of adopting AI. Also, the phenomena of organizational development within the construction industry may impact on the adoption of AI Technology was found to be significantItem Factors facilitating conversational artificial intelligence at an insurance organisation in Johannesburg(2022) Tshifularo, Setebatebe BGlobally the insurance industry is changing significantly due to new technological advances, data growth asymmetry and more recently the Covid-19 related disruptions (Balasubramanian, Libarikian, & McElhaney, 2021). Further to these changes, are consumer tastes and preferences fuelled by availability of information and accessibility of platforms such as social media. Consumers have become more price and product sensitive due to the many options available from competing products and companies. To keep up with these changes which present both risks and opportunities, insurers have been long exploring with technologies such as Artificial Intelligence, Internet of Things, Cloud computing to name a few (Santenac, Majkowski, Manchester, & Peters, 2019). Artificial Intelligence is a buzzword in the insurance industry as most insurers are looking at ways to drive effective distribution channels, be cost effective and exceed customer expectations (Accenture, 2018). Conversational AI is a subset of AI which this study seeks to explore in depth. The aim is to consider global, continental, regional and local insurers’ implementations of conversational AI and derive insights on factors that led to successful CAIs implementations.