Exploring ChatGPT usage amongst female academics and researchers in the academia Open Access

Artificial intelligence (AI) has emerged as a transformative force globally across various sectors (Adiguzel et al., 2023), including education. A report released by (UNESCO IESALC) in 2021 revealed that there is a substantial stride in the enrolment of women in higher institutions while at the same time indicating inequality between men and women in many areas, including research performance (Unterhalter et al., 2022).

Like other African countries, women in South Africa have historically faced numerous challenges in their academic careers. These challenges include gender bias, unequal access to resources, and work-life balance issues (Mdlenleni et al., 2021), hindering their research productivity and career advancement. South Africa has made substantial progress in promoting gender equality in higher education, yet disparities persist. The statistics from the report released by the Department of Higher Education and Training (2023) revealed that women constitute 48.7% of academic staff in South African universities. However, they are underrepresented in senior academic positions and research leadership roles. Their underrepresentation is evidenced in research output, where female academics often publish fewer papers than their male counterparts (Mouton et al., 2022). In a recent study by Nygaard et al. (2022), evidence showed that men outperformed women in academic publications, with men having 6.81 while women having 4.29 means. This performance has implications for assigning leadership responsibility in many African institutions (Adewale and Potokri, 2023a), academic promotion, and career advancement (Beaudry et al., 2023).

The emergence of Artificial Intelligence technologies offers a promising panacea to these disparities by providing tools and resources that can support and enhance the research and academic activities of female academics and researchers in the ivory towers in South Africa. Numerous AI-powered platforms can ease literature reviews, research data analysis, and research dissemination, significantly reducing the time and effort required for the research tasks. These AI tools for literature reviews and refining ideas like ChatGPT and Research Rabbit can help researchers swiftly identify relevant studies, thereby accelerating the initial phases of research (Remadevi and Arunkumar, 2023).

Moreover, the use of AI can help address the structural inequalities in career advancement and opportunities in academia (Santos et al., 2021) and the issue of work-life balance, which is one of the significant barriers for many female academics. AI enables researchers to concentrate more on creative and high-level aspects of their work by automating routine research tasks such as data entry and analysis. This will improve their efficiency and provide more flexibility, allowing female academics to manage their professional and personal responsibilities better (Chiu, 2023; Zhang and Aslan, 2021).

The impact of AI on research productivity is evident in various metrics. Wee (2024) reported the outcome of a study conducted by McKinsey (2020), which found that organisations leveraging AI for research and development experienced a 30% increase in research output. Although specific statistics for South African tertiary institutions are limited, similar trends are anticipated, given the applicability of AI. In a recent study, Burger et al. (2023) used AI to support their literature search for the systematic literature review. Chubb et al. (2022) noted that significant funders of academic research now consider AI’s role in conducting and evaluating research. At the same time, there is concern that users will maintain research integrity. It should be reiterated that AI can foster the spirit of collaboration and networking for research productivity. Female academics facing networking barriers in traditional academic settings can leverage AI-powered platforms like ResearchGate to connect other researchers with similar research interests, facilitating collaboration (Chubb et al., 2022; Esangbedo et al., 2024).

The rapid advancement of Artificial Intelligence technologies has significantly impacted all fields, including academia. One such AI tool, ChatGPT, has attracted attention for its potential to support academic work through drafting, editing, and information retrieval tasks. Despite its potential benefits, there is limited empirical evidence on adopting and using ChatGPT among female academics and researchers, particularly in South Africa. Previous studies have highlighted the transformative potential of AI tools in the academic arena. It has been found in their study that AI have a vast potential to connect knowledge and knowledge producers, providing personalised and supporting administrative tasks (Bahrini et al., 2023; Chen et al., 2020; Chubb et al., 2022; Kooli, 2023; Rudolph et al., 2023). Despite being in its early stage, it has been proven that AI can positively and productively enhance academic research as it reduces the burden on researchers and accelerates the research pace (Gandhi and Gandhi, 2023). These tools offer researchers powerful means to enhance productivity, access information, and engage innovatively with research enquiries (Adiguzel et al., 2023). However, these studies did not delineate data by gender, leaving a gap in understanding how female academics, who may face unique challenges in their professional circle, engage with AI tools like ChatGPT. Therefore, this study seeks to fill this gap by examining the motivations, patterns, benefits, and challenges associated with ChatGPT usage among female scholars in South Africa.

The following research questions are raised to guide this study:

• (1)

  What are the primary motivations behind the usage of ChatGPT among female researchers in South Africa?

• (2)

  How do female researchers in South Africa perceive ChatGPT’s impact on their academic work and career advancement?

• (3)

  What challenges do female researchers face in South Africa when using ChatGPT for academic support?

• (4)

  What strategies can be identified to maximise the effectiveness of ChatGPT in bridging the academic gender gap for female researchers in South Africa?

The gender gap in academia remains pervasive, with women often underrepresented in higher education, research, and leadership positions (Van den Besselaar and Sandström, 2016). Like many other countries, South Africa has not been immune to these disparities (Mdlenleni et al., 2021). Factors contributing to this gap include implicit bias, the digital divide (OECD, 2018), limited mentorship opportunities, family and work-life balance challenges (Ojobanikan and Potokri, 2022) and women’s preference for teaching over research (Mouton et al., 2022) and the outbreak of COVID-19 pandemic (Madsen et al., 2022). Using AI technologies such as ChatGPT presents a promising avenue to mitigate some of these obstacles.

In recent years, efforts to address gender disparities in academia have gained momentum worldwide (Abdulkadri et al., 2022), focussing on understanding female researchers' unique challenges. One emerging avenue in this pursuit is using artificial intelligence, such as ChatGPT, to bridge the academic gender gap. This literature review aims to provide an overview of the existing research landscape related to ChatGPT usage among female researchers in South Africa, shedding light on the potential of AI-driven support in academia.

Globally, research outcomes have continued to highlight a wide gap between male and female academics in research outputs. It has been empirically confirmed that no continent is spare from this challenge as male academics have more research productivity than females (Abramo et al., 2021; Nygaard et al., 2022). Also, regarding citations on research output, women still score lower than men (Van den Besselaar and Sandström, 2016). Despite this undisputed disparity based on gender, (Ma et al., 2023) submitted that reducing the disparity in research output between males and females has been an arduous task for researchers. It is therefore essential to devise a working means of closing the in line with the Sustainable Development Goal (SDG) 5. Studies have attributed this to different factors, among which are balancing family responsibilities and research work (Abramo et al., 2021; Van den Besselaar and Sandström, 2016), propensity for both international and local collaboration with women having low performance (Abramo et al., 2019). Given this, there is a need for an everlasting solution to address this. There still needs to be a gap in how female researchers leverage technology to catch up. Mouton et al. (2022) recommended more interventions to enhance the research performance of females in South Africa. Menon and Shilpa (2023) suggested that engagement of researchers with Chatbot will go a long way in improving research productivity. In South Africa, this apparent gender disparity in research productivity has impacted academia.

The evolution of ChatGPT in December 2022 by Open AI (Dai et al., 2023; Rosenblatt, 2023) mirrors the broader trajectory of AI, a journey from humble beginnings to a transformative technology that influences academia, business, healthcare, and beyond. It exemplifies the ongoing quest to create machines that mimic human intelligence (Lin, 2023) and augment our capabilities, making our interactions with technology more natural and productive. As AI advances, we can only anticipate further breakthroughs and innovations that will shape the future of human-computer interaction (Lin, 2023).

The impacts of AI, including ChatGPT, on academics and researchers have been the subject of extensive research and discussion. There has, therefore, been a mixed reaction to the impacts of ChatGPT among scholars (Dai et al., 2023). Some saw it as a blessing promoting efficiency and effectiveness, while others perceived it as a retrogressing tool for academics and researchers (Menon and Shilpa, 2023). Whichever way or position one maintains, Higgins (2018) has cautioned on the use of artificial intelligence thus: “It will either be the best thing that’s ever happened to us, or it will be the worst thing. If we are not careful, it very well may be the last thing”. This calls researchers' attention to the project and envisages the likely positive and negative impact that the use of artificial intelligence may have on academics and researchers in academia.

Artificial Intelligence promises to provide education and training and enhance human performance, improving proficiency in various tasks and activities (Menon and Shilpa, 2023; Yang, 2021). For instance, ChatGPT has been helpful to researchers writing academic content. However, this has been criticised and declared ethical because its accuracy is questionable (Chen, 2022). Artificial intelligence (ChatGPT) has enabled the development and implementation of systems that serve as highly effective educational aids (Chen et al., 2020; Kaid et al., 2023). The tool can assist academics in many ways, such as helping generate assessment questions, grade examinations, and analyse them. Another impact of ChatGPT on its users is the ability to provide real-time assistance (Dai et al., 2023). This is demonstrated in its ability to provide unique responses to its users coherently and satisfactorily. It helps researchers address difficulties in report writing in a textual manner of different styles that suit what is needed at that moment (Bahrini et al., 2023).

Rahman et al. (2023) conducted an experiment to determine how impactful ChatGPT use can be to researchers, and it was found that it is effective in generating valuable ideas for researchers and brainstorming. However, the onus is on individual users to ensure the generated ideas are cross-checked for accuracy. ChatGPT serves many purposes (Bahrini et al., 2023), ranging from aiding in your learning, coding, analysis, and writing endeavours to support your teaching, emotional well-being, and job application processes (Lin, 2023).

Leveraging ChatGPT can transform academic activities into efficient, time-saving endeavours (Eke, 2023). By incorporating ChatGPT into various aspects of academia, researchers and students can unlock many benefits (Bahrini et al., 2023; Kooli, 2023). ChatGPT can be a valuable research companion, aiding researchers in literature reviews, data analysis, and hypothesis generation. Its ability to swiftly scan through vast databases of academic literature and provide concise summaries can significantly expedite the research process. ChatGPT is considered promising for scientific research (Sallam, 2023) because it helps in many circumstances in writing the research report and conducting other parts of the research. Since ChatGPT cuts across all disciplines, including academic research, researchers must be meticulous. This could be achieved by conflating the output from ChatGPT and their human analytical acumen. Malik et al. (2023) emphasised that during the research process, it is essential to recognise that ChatGPT should not be a substitute for the necessity of analytical and imaginative thinking among researchers.

Envisioning the negative impact of the use of ChatGPT in academia, Lyle submitted that “while the tool may be able to provide quick and easy answers to questions, it does not build critical thinking and problem-solving skills, which are essential for academic and lifelong success” (Rosenblatt, 2023). However, it must be mentioned that the developer of this tool acknowledged that it has some limitations (Eke, 2023). Also, the tool can only uphold academic integrity if ethical issues are observed. One way of undermining academic integrity is when academics, researchers and student alike contract their academic writings to ChatGPT (Eke, 2023) because they see the tool as capable of giving some correct output.

Generally, AI is penetrating every aspect of human lives. It is filtering, which can be felt in both personal and professional aspects of human lives, and is not all yielding a positive impact (Borenstein and Howard, 2021). For instance, artificial intelligence use may erode the academic integrity of academics and professionals if not ethically related. Even before ChatGPT emerged, there were academic fraud cases, such as plagiarism. The fear of the negative impact and unethical use of ChatGPT necessitated the announcement of a ban on it in the United States (Rosenblatt, 2023). In addition to the lack of transparency concerning content generation, ethical concerns, particularly the potential for bias rooted in training datasets and issues related to plagiarism, were recurrently brought up (Sallam, 2023). These ethical dilemmas pose significant challenges when relying on AI models like ChatGPT to generate content in academic and research contexts.

The risk of bias in AI-generated content is paramount, as the models are trained on vast datasets that may reflect and perpetuate societal biases. This bias can manifest in skewed perspectives, favouring specific demographics or viewpoints and undermining the integrity of research and academic work (Sallam, 2023). Furthermore, the risk of unintentional plagiarism is critical, as AI-generated content may inadvertently reproduce existing material without proper attribution, potentially leading to ethical violations and academic misconduct.

Addressing these ethical concerns is pivotal in ensuring that AI tools like ChatGPT are used responsibly and ethically within the academic community (Bahrini et al., 2023; Kooli, 2023). It calls for transparent disclosure of AI assistance in research and writing, as well as diligent oversight to minimise bias and plagiarism risks, thereby upholding the highest standards of integrity in scholarly work.

Despite the beauty and numerous benefits of using ChatGPT, studies have also shown its challenges and limitations (Adiguzel et al., 2023; Sallam, 2023). One of the challenges to its adoption in the education system is that it can widen the gaps in existing inequalities (Adiguzel et al., 2023). Numerous users have consistently highlighted that one of the primary limitations they encounter when employing ChatGPT for scientific writing is the prevalence of superficial, inaccurate, or erroneous content (Sallam, 2023). This drawback has raised concerns about the tool’s reliability and suitability for generating precise and trustworthy scientific content. Researchers often need to verify and extensively revise the AI-generated text to ensure its accuracy and appropriateness for scholarly purposes. As a result, while ChatGPT offers valuable support in various writing tasks, its limitations in producing scientifically rigorous and reliable content underscore the importance of human review and critical evaluation in academics (Kooli, 2023) and research settings. ChatGPT’s limitations also include its incapability to comprehend the shared content and its inability to assess the pertinence or correctness of the information (Rudolph et al., 2023). This is because it functions solely as a text-generating machine, creating a reasonable facsimile of the information it is provided without the ability to understand or evaluate it deeply.

In light of the numerous limitations and ethical considerations around using ChatGPT and other AI in academia, concerted efforts must be made (Kooli, 2023). This can be done by providing AI education and awareness for female academics and researchers, especially in academia. They should clearly understand what ChatGPT can and cannot do. Recognising that ChatGPT is a machine learning model and may provide inaccurate or incomplete information is crucial. Being aware of its limitations is the first step in managing them.

This study adopted a qualitative research design of phenomenology type to explore the usage of ChatGPT among female academics and researchers. This aligns with Creswell (2014) and Adewale and Potokri (2023b), who submit that qualitative research allows for an in-depth exploration of the object of study. “quantitative researchers study things in their natural settings, attempting to make sense of, or interpret, phenomena in terms of the meanings people bring to them” (Aspers and Corte, 2019, p. 142)

This study is a case study which was conducted using one selected University in South Africa. The selected institution shares feature similar to other tertiary institutions in South Africa. It is a co-educational university with male and female researchers and has 21st-century technological facilities like the internet, electricity, and a well-equipped library that researchers can use. The university is uniquely promoting the Fourth Industrial Revolution (4IR) ideas.

This study sampled 20 female academics and researchers using purposive and convenient sampling techniques. The purposive sampling technique enables researchers to select only the participants who possess the qualities suitable for the study (Uleanya, 2023). Specific inclusion criteria identified by the researcher determined the selection of 20 participants in this study. These inclusion criteria include being a female, a researcher, and an academic member of the purposively selected university in South Africa. The researcher believes that the selected participants would be able to proffer answers to the research questions raised to guide the study. Thus, this study included four senior lecturers, nine postdoctoral fellows, and seven doctoral students. Semi-structured interviews were conducted to elicit data from each of the participants.

The sample was limited to 20 when the saturation was reached in the data generated through the semi-structured interviews conducted.

The researcher prepared an interview protocol based on the study’s research questions to guide the conduct of the unstructured interview sessions. The choice of semi-structured interview was to give room for follow-up questions where and when necessary, during the interviews and to clear ambiguities in the participant’s responses. The interview was recorded using an audio gadget with the consent of the participants.

Upon the interview’s conclusion, the researcher used Microsoft 365 to transcribe the audio data by converting it into a text document. Afterwards, the produced (transcribed) text was proofread while listening to the audio record for accuracy and confirmation. The text document was imported into NVivo software for qualitative coding and categorisation into themes and sub-themes. Qualitative coding is when qualitative data are systematically analysed by assigning labels or codes to segments of information based on themes, patterns, or concepts (Brailas et al., 2023; Wolff et al., 2024). On the other hand, thematic analysis enables researchers to systematically organise and present complex datasets meaningfully (Dawadi, 2020). In other words, it helps organise, sort, and interpret data to identify recurring ideas or categories within the dataset, facilitating the extraction of meaningful insights. The use of thematic analysis in this study helped to summarise critical features of a large dataset and produce an organised report (Nowell et al., 2017). Meanwhile, the participants' demographic information was analysed with descriptive statistics using frequency counts and percentages. The researcher also used verbatim questions from the participants to lend their voice to the findings.

The issue of ethics in conducting research was strictly adhered to by the researcher in conducting this study (Alase, 2017). The researcher ensured the informed consent of the participants in writing, detailing the study’s purpose and roles (Manti and Licari, 2018). Their participation was voluntary, meaning they were not coerced and could withdraw their participation at any stage. In addition, the participants were assured of the confidentiality of the information provided and that their real identities would not be revealed, upholding the principle of anonymity in research (Kang and Hwang, 2023). Furthermore, the researcher specifically sought the participants' permission to record the interview on the audio gadget. Kural (2020) submitted that “Recording does not only help you retain the information you find relevant at an interview but also gives you a chance to listen and discover new themes and answers you have not thought of during the interview (par. 7).”

The researcher adopted the Trustworthiness, Auditability, Credibility and Transferability (TACT) framework (Daniel, 2019) to ensure the quality and reliability of this qualitative research. Lincoln and Guba (1985) made the expository presentation in 1985 to demonstrate that the assessment of research trustworthiness, which comprises credibility, transferability, dependability, and confirmability, should be used to measure the validity and reliability of the qualitative study and should not be determined by scientific measures.

Trustworthiness involves confidence in the data sources, research outcome and researcher’s experience. The most important aspect of trustworthiness in qualitative research is transparency because it determines the relevancy of any research outcome (Adler, 2022). This study achieved this by familiarising oneself with the participant and ensuring neutrality in reporting the outcome. The researcher did not add or reduce the findings or allow personal views to override the study’s outcome (Daniel, 2019). Daniel (2019) further asserts that Auditability “requires the application of systematic procedure for collecting, analysing and interpreting data (page 120).” In this study, the researcher systematically interviewed, fulfilling the ethical requirement, sought the participants' consent to interview and record on the audio gadget, transcribed the data, used NVivo software to code, generated themes and reported the findings.

This section presents the analysis of the collected data in a meaningful and comprehensive manner using tables and figures.

Table 1 provides an overview of the characteristics of the study participants. The table includes their serial number (S/N), academic position, discipline, years of academic experience, familiar AI tools, and experience with ChatGPT. The participants consisted of senior lecturers, postdoctoral research fellows, and doctoral students from various academic disciplines, including Early Childhood Development, Inclusive Education, Science Education, Educational Leadership and Management, law, engineering Counselling Education, Psychology Education, Botany, and Counselling Education, economic, mathematics and statistics—their years of experience in academia range from 1 to 12 years.

Regarding familiarity with AI tools, participants reported using tools such as Scholarcy, Quill Bot, Grammarly, Elicit, Scispace, Scopus AI, Word Tune, and Consensus AI. Notably, all participants were familiar with ChatGPT, and the majority had prior experience using it. It is important to note that while some participants had extensive experience with AI tools, others had limited or no experience, which may have implications for using these tools in their academic work.

Figure 1 presents the motivations driving the utilisation of ChatGPT among female researchers in South Africa. As shown in Figure 1, seven reasons were identified from the interview: the drivers to use ChatGPT ease and convenience, free access to its contents, the need to reduce workload, literature review assistance, idea generation, and multilingual support. One participant reported that I sometimes face challenges in setting higher-order questions, particularly in computational thinking, which is a new field within the context of my teaching practice. I need to understand better the kinds of questions I ask my students.

Another participant revealed her motivation thus I often faced challenges in getting started with writing assignments. ChatGPT provided an initial draft or structure, making beginning and maintaining momentum easier.

Figure 2 revealed the impact of ChatGPT on females' academic work and research in tertiary institutions in South Africa. The impacts were categorised into two main categories: positive and negative impacts. The Positive impacts identified are help in time management, writing improvement, improved productivity, and access to skills development content.

One of the participants sharing some positive impacts of using ChatGPT says,…I have used it for academic purposes and find it quite helpful. My experience with it is that it is Very, very convenient. It makes you work faster. It helps one to develop content. Writing a kind of sentence or statement for tragedy will help you restructure it, expand on it, reframe it …

Another participant reported the impact of ChatGPT on her that the significant benefit of using Chat GPT is that it has helped enhance my productivity in setting exam questions. My writing skills have also improved.

On the other hand, perceived negative impacts are incorrect suggestions or contents, erosion of human critical thinking, prejudiced positions, procrastination of assignments, a boycott of conventional libraries, and it encourages complicity. On the negative impact of ChatGPT, another participant said: … I have concerns about the ethical implications and biases present in AI models like ChatGPT. I am worried about reinforcing existing biases or the responsible use of ChatGPT in research and academic contexts …

Table 2 revealed that 11 challenges were identified as limitations to using ChatGPT among female academics and researchers in South African tertiary institutions. These are Limited Training, unequal access to technology and its components, low awareness, perceived privacy and security concerns, subjectivity, and limited contextual knowledge. Other barriers are a need for more collaboration among researchers, reliability and validity concerns, plagiarism, and bias; it eliminates social interaction and replaces minds. This finding implies that the challenges faced by female academics and researchers in academia in using artificial intelligence tools like ChatGPT and others transcend academic integrity. One of the participants shared her challenge with using ChatGPT thus … my concerns is about data privacy and the security of research-related information. The confidentiality of my work when using online platforms is not guaranteed.

One of the participants said I do not see any benefit because it denies the students the opportunity to interact with their work. It replaces the mind of a student and the development of critical thinking.

Figure 3 presents the strategies and recommendations for optimising the use of ChatGPT among female researchers in South African Universities. The results of the interviews yielded nine strategies and recommendations. The strategies suggested are awareness creation, mentorship, collaboration between academics and AI developers, localisation of ChatGPT contents, quality control and quality assurance, improved internet access, policy support to allay fear, further training on ethical use of AI, and verification and paraphrasing. One of the participants reported this:

This study aimed to explore the perception of female academics and researchers on the use of ChatGPT in South Africa. This study found that the primary motivation behind the usage of ChatGPT among female researchers in South Africa is that the AI tool has the potential to aid their research productivity. This finding is in congruence with the outcome of the study conducted to experiment with ChatGPT by Rahman et al. (2023). It was confirmed that the tool is capable of generating new research ideas. Sometimes, research gets bereaved of ideas on what to research. In their submission, Bahrini et al. (2023) found that ChatGPT can assist researchers in generating research ideas, writing research methodologies and performing statistical analysis.

Furthermore, the study’s outcome revealed that female researchers in South Africa’s perception of ChatGPT’s impact on their academic work and career advancement yielded positive and negative impacts. While some thought that the AI tool has made research more exciting and manageable, others had negative perceptions, feeling that it has made them lazy and unable to think independently. This is in congruence with Menon and Shilpa (2023) that ChatGPT is helpful to academics in simplifying tasks and getting tasks done. It also allows for creative writing and content creation (Ray, 2023), which are freely accessible to all, irrespective of gender, race and affiliation.

The study has also found that the effort to use ChatGPT by female researchers in South Africa for academic support has some obstacles, including a lack of comprehensive digital technology knowledge and a lack of collaborations. This is in agreement with the submission of Eke (2023) that some institutions of learning, especially in the developing world where poverty is well pronounced, need to have digital technology awareness and knowledge and have access to technology. In another study, Hyun and Kim (2023) found that creepiness, which is the discomfort experienced by researchers when relating with an unknown person, is a challenge to using ChatGPT. As such, it aggravates researchers' fear, especially female gender and ultimately not wanting to use AI for their academic works. These findings are in congruence with the experiment conducted to determine the accuracy of ChatGPT. Rahman et al. (2023) found that the tool confessed that it could not access the source. This means that the use of AI tools like ChatGPT should be independent of it as it has many conspicuous imperfections in the responses it gives to its users.

Addressing ethical concerns is paramount when utilising AI tools like ChatGPT in academic and research settings. Two key strategies for mitigating these concerns include using AI detectors and by using declarations and disclaimers. The implementation of AI detectors, which are specialised software or algorithms designed to identify potential ethical issues, such as bias, plagiarism, or inappropriate content, in AI-generated output. These detectors can scan the content produced by AI tools like ChatGPT and flag any bias, plagiarism, or other ethical breaches. Appropriate corrective actions can be taken upon detection, including revising the content, providing proper attribution, or conducting further review to ensure ethical compliance.

They also incorporate declarations and disclaimers in research papers, publications, or documents generated with the assistance of AI tools. These declarations can transparently acknowledge the use of AI technology in content creation and inform readers about the potential limitations and ethical considerations associated with AI-generated content. Disclaimers can emphasise that readers and researchers should critically evaluate the AI-generated content and that the responsibility for verifying accuracy and ethical compliance lies with the human user. Such declarations and disclaimers uphold transparency and integrity in research and academic work while raising awareness about AI’s role in content generation. Menon and Shilpa (2023) found that users of ChatGPT expressed their scepticism about the accuracy of the information provided by the AI despite its speed and humongous advantages.

Lastly, the study found awareness creation, training, and collaboration between academics and AI developers, among others, to be strategies that can be employed to maximise the effectiveness of ChatGPT usage in bridging the academic gender gap for female researchers in South Africa. This finding is congruent with Eke (2023), who found ChatGPT technology beneficial to research and teaching. This presupposes that the tool is capable of making university service delivery realistic. Montenegro-Rueda et al. (2023) also mentioned the role of ChatGPT in engaging individuals to dwell more in personalised learning and feedback, which is capable of developing potential. At the same time, it also allows for collaboration, eventually building virile teamwork among academics.

Conclusively, this study examined the perceptions of female academics and researchers South Africa concerning the use of ChatGPT, revealing both its potential to enhance research productivity and the challenges it poses. While many found the AI tool helpful for generating new research ideas, methodologies and performing statistical analysis, others expressed concerns about its impact on independent thinking and academic rigour. The study identified the significance of addressing digital literacy, encouraging collaboration, and implementing ethical measures to combat potential inadequacies. Therefore, raising awareness, providing training, and promoting partnership between academics, institution leadership and AI developers are critical strategies to maximise the benefits of ChatGPT and bridge the academic gender gap.

This study contributes to the technology acceptance model (TAM) by illustrating how perceived ease of use and perceived usefulness influence the adoption of ChatGPT among female researchers. It indicates that perceived productivity enhancement is a potential motivator. In addition, both positive and negative perceptions emanating from the study contribute to the theoretical understanding of the impacts of AI in academic works. More importantly, the findings of this study added a gendered perspective to the discourse on AI adoption in academia, highlighting the unique experiences and challenges female researchers face. This can inform theories related to gender and technology, underscoring the need for more inclusive and supportive digital environment.

Since this study has revealed that leveraging AI like ChatGPT enhances research productivity among female academics, institutions should organise training programmes for staff members to sharpen their technology knowledge. These training programmes and providing access to digital technology will bridge the digital divide based on gender. Management of various academic institutions must foster mentorship programmes that connect experienced researchers with early-career female researchers to guide AI tool integration. As a matter of urgency, the government should customise AI tool content to align with South African research needs, including support for local languages and contexts. The leadership of tertiary institutions should advocate for institutional policies that support the equitable use of AI tools and technology access for female researchers. Similarly, institutions should embrace the incorporation of AI detectors and ethical disclaimers to ensure the responsible use of AI in research. Above all, a should be created in the tertiary institutions to encourage collaboration between academics and AI developers to enhance the effective use of AI tools in research.

This research effort was delimited to a university in the urban area of South Africa, among the academics and the use of ChatGPT out of numerous AI tools. This implies that the findings may only be generalisable to some populations outside the target population. Similarly, qualitative research design is often criticised for being subjective. Hence, further research on AI tool usage among other groups who may not be in academia should be conducted. Also, the subjectivity of the qualitative method could be reduced, if not eliminated, using both Mixed and quantitative research designs.