The Challenge of Adopting AI in Southern African Higher Education: Media Educators’ Perceptions

Abstract

This article explores perceptions of educators on challenges of integrating AI in journalism education. It draws on Bourdieu’s field theory, especially the idea of external shocks, social shaping of technology theory and technological domestication theory. Data were gathered through interviews with selected lecturers from universities and colleges in Zimbabwe, Namibia, South Africa and Botswana. The findings show that educators in Zimbabwe and Botswana viewed AI as compromising the integrity of journalism education and as having a psychological toll on them. Contrastingly, educators in Namibia and South Africa, while admitting AI’s shortcomings, emphasised its potential to enhance journalism training in Africa.

1. Introduction

Every technology represents a significant disruption, for better or worse, of the status quo. How communities embrace these technologies is a function of the technologies’ own affordances, the public discourses around which this technology is framed and users’ own experiences and perceptions of the technology. Artificial Intelligence (AI) is one such technology that has disrupted society and divided public opinion, even at a global level. Recently, there has been a surge in discussions around the use of artificial intelligence in education. These discussions have been spurred on by recent new AI tools such as ChatGPT and Deep Seek (Chen et al., 2020). But in Africa, the discussion of the adoption of AI in education has a short history. Ncube et al. (2025) note that the adoption of AI in journalism training in Southern Africa is at a nascent stage because the curricula of Journalism Schools (J-Schools) betray a lack of courses dedicated to the subject of Artificial Intelligence. This could be because of lack of the requisite infrastructure in the universities. As UNESCO (2019, p. 5) states, “AI requires advanced infrastructures and ecosystem of thriving innovators”. Coming from this background, this article interrogates the challenges and opportunities that the adoption of Artificial Intelligence (AI) in journalism and media education presents to the university and college journalism and media educators in Southern Africa. Specifically, the study focuses on selected Zimbabwean, South African, Namibian and Botswana journalism and media educators. The article seeks to answer the question: how do university educators in African higher education perceive of AI’s entrance into their field? Previous research (see, for example, Ncube et al., 2025) has focused on how journalism educators are ‘recurriculating’ and integrating AI as part of the journalism curriculum. This extant research has not focused on how AI is being integrated as a teaching tool by journalism educators. This research seeks to close this lacuna by focusing on journalism educators’ perception of integrating AI tools for journalism education. Our argument is threefold. First, we argue that emerging AI technologies in higher education are increasingly becoming sources of mental/psychological strain. Second, AI technologies are etched in hazy ethical controversies that African academics rarely understand how they work. Third, to understand AI adoption in African contexts, we need to understand the disruptive power relations and tensions that are emerging and shaping the field.

2. AI Adoption in Higher Education—A Literature Review

As AI technologies literally erupt in society (Singer, 2024) and drive communities into algorithmic societies (Singer, 2024), it is important to understand how specific communities of practice, like journalism educators, are challenged by AI and the opportunities it potentially presents in education. In other contexts, research has shown that AI’s application in education presents both opportunities and challenges (Ouyang & Jiao, 2021; UNESCO, 2019). Scholars argue that AI has “produced an opportunity for the reform of teaching and learning” (Huang et al., 2021, p. 207). AI in education has also been noted to provide new ideas and solutions for teaching and learning; efficiency and quality; and personalised learning to address individual learners’ needs (Huang et al., 2021; Chen et al., 2020; UNESCO, 2019). The ability of artificial intelligence “to support the engagement of learners with disabilities” (Chen et al., 2020, p. 40) and to reduce the burden on the teacher (Huang et al., 2021; Chen et al., 2020) has also been hailed.

However, a multiplicity of challenges associated with the adoption of AI in education have also been noted. First, scholars have argued that there are ethical issues associated with the adoption of AI in education (UNESCO, 2019). Second, the development of AI policies in developing countries is at a nascent stage (UNESCO, 2019). Third, AI in education may further widen digital inequalities by creating new forms of digital divide whilst the new AI-driven teaching methods “collide with mainstream traditional practices often without rigorous evaluations supporting the claimed benefits of new solutions” (UNESCO, 2019, p. 28). Lastly, it is mentioned in Chen et al. (2020, p. 1) that AI in education challenges “the instructor’s role…”. This study acknowledges the foregoing studies and builds on them. Specifically, it seeks to interrogate if the challenges and opportunities noted in non-Sub-Saharan African educational contexts are similar to those faced in Southern African educational contexts—contexts with different political and economic systems. Whereas UNESCO (2019) sought to examine challenges faced in the global south, this article argues that the global south is too broad with varied educational, political, economic and socio-cultural systems and each deserves to be studied in its own right and not be bunched together. In light of the foregoing, the study seeks to answer the following questions: how do university journalism educators in African higher education perceive AI’s entrance into their field? How are the educators adapting to, and adopting, AI in higher education? What are the opportunities and challenges presented by AI to journalism education in Southern Africa? The question of how users perceive AI has also garnered varied responses. However, in this growing body of literature, there is a dearth of studies that interrogate AI’s adoption in journalism education, as well as its impacts, opportunities and challenges.

3. Conceptual Framework

This study deploys an eclectic conceptual framework fusing Bourdieu’s theory of practice, social shaping of technology, and technological domestication theory. Such analytical insights are critical in examining the perceptions of Southern African journalism educators on the integration of AI in the journalism teaching and learning ecosystem. Bourdieu’s theory blends a range of philosophical arguments, styles, and accounts of social relations fusing structural and lived experiences into an epistemologically grounded mode of inquiry that is universally applicable yet distinct (Ndu, 2022). Bourdieu’s theory of practice is constituted by four intricately intertwined concepts: practice, habitus, field and capital (Bourdieu, 1990). The concepts must not be looked at in isolation but as an embodiment of essential elements that propel the other into existence (Munoriyarwa & Chiumbu, 2019; Ndu, 2022). The aspects are critical in our exploration of the meanings attached to AI in journalism training by educators. Habitus refers to entrenched habits, skills and dispositions. In the context of this present work, habitus may be taken to mean what UNESCO (2019, p. 28) refers to as “mainstream traditional [teaching] practices”. In other words, habitus is the dispositions that people have accumulated throughout their lives, as expressed through their attitudes and behaviours towards issues (Munoriyarwa & Chiumbu, 2019). Habitus is critical in exploring embodied beliefs, aspirations and socially constructed habits that influence meanings attached to the disruptive impacts of AI in journalism training. The field concept views society as being composed of different but interconnected fields, with each field having its specific logic (Munoriyarwa & Chiumbu, 2019). Critically, the field is a space where perspectives are oriented and where habitus thrives and strives to exist in tandem with other concepts with which it interacts (Bourdieu & Wacquant, 1989). We consider journalism training as a field currently experiencing the disruptive impacts of AI, and consequently presenting an array of challenges and opportunities. The field’s structure is historically located in its past struggles and challenges, resulting in established values (Bourdieu, 1990). We also illuminate on the struggles and politics characterising the journalism training field in Southern Africa and implications for educators and learners. We also examine how Southern African journalism educators cope with the disruptive impacts of AI in the training process.

The field concept also helps to explain forces, both internal and external, that shape and structure the perceptions of Southern African journalism educators on the integration of AI in journalism teaching and learning. Critically, Bourdieu’s concept of external factors and internal factors is relevant to the study. AI can arguably be viewed as part and parcel of what Bourdieu terms external shocks, which disrupt the power relations in the field of journalism training. Such shocks have the ability to reshape journalistic perceptions and practices because the fields are structured by power relations, with various forms of capital being valued differently across each field.

Since Bourdieu’s theory does not specifically focus on technological innovations, we appeal to the social shaping of technology theory (SST) in order to complement its limitations. The social shaping of technology theory is viewed as a critique of McLuhan’s technological determinism theory. There is no agreed definition of what constitutes an SST approach (Russell & Williams, 2002; Howcroft & Taylor, 2022), but it is generally viewed as a broad theory that encompasses many traits (R. Williams & Edge, 1996). This is because this tradition is constituted by diverse disciplines and intellectual origins (Howcroft & Taylor, 2022). Importantly, SST theorists underscore the discourse of shaping rather than social construction of technology (MacKenzie & Wajcman, 1999). The SST tradition argues that technology and society are mutually constitutive, rather than separate spheres (MacKenzie & Wajcman, 1999). In fact, society and technology influence each other interdependently and in multifaceted ways rather than linearly (Howcroft & Taylor, 2022). Furthermore, the SST submits that technology is used and appropriated in different ways depending on users’ political, economic, social and individual predilections (Pobuda, 2022). In this regard, we examine how perceptions of journalism educators across Southern Africa are informed and shaped by their political, social and individual predilections.

To make our conceptual analysis more robust, we also borrow insights from the technological domestication theory. This theory is commonly used in multidisciplinary studies examining the integration of technologies in everyday life (Fortunati, 2009; Nimrod & Edan, 2022). The theory illuminates on adoption, rejection and use of new technology within households (Silverstone et al., 1992). It considers the sociocultural landscape within which technology adoption occurs and covers the pre- and post-adoption stages (Haig-Smith & Tanner, 2016). Critically, the theory avers that when users engage with technology, four intersecting processes occur: appropriation, objectification, integration and conversion (Silverstone et al., 1992; Nimrod & Edan, 2022). Domestication theory further notes that integrating new technologies can create new burdens, tension and symmetries that necessitate further negotiations (Wang et al., 2024). Domestication of technologies, furthermore, is about how users appropriate, objectify, incorporate and convert technologies on an everyday basis (Silverstone & Haddon, 1996). Originally, the theory was developed to understand the uptake of technologies at a family/individual level (Wang et al., 2024, Silverstone & Haddon, 1996).

In this paper, the domestication theory of technology helps us understand how individual users define and shape how these technologies are used in their specific contexts. Because the domestication theory eschews technological determinism, it allows us to bring it into conversation with the field theory. As we noted, the field concept also helps to explain forces, both internal and external, that shape and structure the perceptions of Southern African journalism educators on the integration of AI in journalism teaching and learning. Domestication theory allows us to go beyond seeing technology as a disruption, but to examine the complexity of AI adoption by journalism educators. Furthermore, bringing into conversation the domestication theory and the field theory allows us to understand the sense-making processes that happen at the multiple levels of negotiation that happen in the process of adopting a technology like AI. We then bring the field theory and the domestication theory into conversation with the SST in order to understand how educators make meaning of AI, and how they perceive AI technologies’ influence in journalism education.

4. Research Methodology

This research is qualitative and interpretive. It draws on interviews with lecturers who were purposively sampled from universities in four southern African countries of Botswana, Namibia, South Africa and Zimbabwe. The researchers purposively and conveniently sampled journalism lecturers. The criterion for sampling was their being journalism lecturers (purposive sampling) and according to their availability (convenience sampling). The use of semi-structured interview questions allowed for open-ended discussions about the ways in which AI technologies are being integrated in journalism education, and the corresponding challenges and opportunities. Verbal informed consent was obtained from participants before interviews commenced. Our interview demographics are presented in Table 1 below.

Table 1. Interview demographics.

Country	Number of Interviews	Type of Interviews
Botswana	5	Face-to-face interviews
Namibia	4	Telephone/face-to-face interviews
South Africa	8	Telephone interviews
Zimbabwe	8	Telephone interviews

In all interviews, we also preferred more ‘personalised’ questioning where we explored interviewees’ own perspectives and opinions about AI in education. Integrating a technology like AI is often a personalised decision, even though, in some instances, institutions and organisations may force it on their employees. Our question revolved around, amongst other issues, the extent to which journalism lecturers are adopting AI. Furthermore, we narrowed down to the challenges they encounter and the opportunities that they perceive as being presented by AI. In total, we had 25 interviewees from seven institutions of higher learning: six universities and one B.Tech-offering polytechnic college. While more respondents could have been added, we believe our sample is sufficient to draw valuable insights into lecturers’ perceptions of challenges associated with integration of AI in journalism education, since we do not seek representation but depth. The respondents were anonymised and differentiated using letters and numbers, e.g., LZ1, LZ2, etc., for Zimbabwean; LN1, LN2, etc., for Namibian; and LS1, LS2, etc., for South African educators.

The following sections present the findings on the lecturers’ perceptions on the challenges and opportunities associated with adoption of AI in journalism and media education in Southern Africa.

5. AI as Cause of Psychological Strain Amongst Educators

The findings demonstrate that AI is causing psychological strain amongst journalism educators. This was both explicitly stated and implied by respondents. The psychological strain is caused by, amongst others, the constant need to catch up with the fast-paced nature of AI-driven changes. As respondent LS2 aptly noted, “…in the world of AI, nine months is an eternity”. Similarly, LZ2 observed the following:

“The rapid pace of AI development can also create confusion about best practices in journalism, making it hard to ensure that students are equipped to navigate these complexities responsibly”. Confusion is a mental or psychological state in which one is bewildered or is not thinking clearly.

The psychological strain is also caused by fear of AI and what it means for their jobs and for education. In other words, educators fear being made redundant by AI through the bastardisation of education, where now they are made to mark machine-generated content and students no longer bother to read and to engage in critical thinking. The threat posed by AI to what UNESCO (2019) calls mainstream traditional teaching practices cause psychological discomfort in lecturers. For example, LS1 asserted the following:

“We are in a department where everyone is amused by AI. Asking questions such as do you know what AI can do? …We are afraid of AI…”. Following on this, it is plausible to argue that fear stems from the knowledge that in the past technology has made some professions redundant and educators are worried that perhaps they are staring at the inevitability of their own profession being made redundant by AI since students can get everything done by machines.

Fear also stems from educators’ concern about the possible misuse or unethical use of AI. For example, respondent LS2 claimed that “…I am very scared about the ethics around it [AI]…”. The same respondent referred to the AI situation as a ‘conundrum’ because, he argued, “See, I don’t call it a problem, although I want to because it’s a problem and an opportunity. So, let’s call it a conundrum”.

The respondent holds conflicting views about AI and this is, arguably, a source of psychological strain. The following are some of the unethical uses of AI that the educators feared: surveillance, mishandling of data, and production and distribution of disinformation and misinformation. For example, respondent LZ4 averred that AI is used as a surveillance tool. He claimed that AI is a tool “designed to harvest information…I still believe AI is a surveillance tool for countries from the South (emphasis ours)” (LZ4, Zimbabwe). With regard to fears around AI’s potential mishandling of data, respondent LN1 noted that “AI does not have the capacity to navigate the complex minefield surrounding ethics and treatment of data”. The fear of fake news has another attendant effect on the lecturers—work overload. As one of the respondents observed, the potential of AI to generate incorrect information breeds a constant need to conscientize students about AI and verify the information. This in turn leads to feelings of being overwhelmed amongst educators. As respondent LN2 noted, the “constant need for verification can feel overwhelming at times”.

It is apparent that the invasion of the journalism and media education field or habitus by AI is causing psychological trauma to the educators. It is clear from the above that it causes the following mental and emotional conditions: fear, confusion and feeling overwhelmed. Being overwhelmed is an emotional feeling that may stem from mental confusion or having a huge workload. Either way, the mind is affected. These psychological issues are a result of AI’s invasion of journalism educators’ field or habitus. The AI is disrupting the traditional practices of teaching and threatening to render the educators redundant. Consequently, the psychological discomfort is manifesting through expressions of fear of AI: confusion about AI, being overwhelmed by AI, doubts over AI’s ability to use data responsibly, fear of decline in students’ critical thinking, etc. These expressed fears and confusion arguably mask deeper psychological turmoil in educators, triggered by fear of being expelled from their habitus by AI. The expressed fears are thus arguably a defence mechanism to deal with the threat posed by AI.

The same is also true of the expressed opportunities; they are equally driven by a desire to co-opt AI in order to not be left behind and be made redundant by it. AI, it is arguable, is causing cognitive dissonance in educators. This cognitive dissonance arises from AI being a set of new technologies, which trigger new beliefs, attitudes and behaviours as well as conflict between journalism educators and their students. Consequently, the individual, for example, in our case, the journalism educator, strives to reduce tension by aligning their thoughts, words and actions (Cancino-Montecinos et al., 2018) or by resisting AI. The above scenario simultaneously demonstrates co-option and resistance. This psychological conflict is best captured by respondent LS2’s characterisation of the situation as a ‘conundrum’, because there are both threats and opportunities. Others characterised it as ‘fear’ and ‘confusion’. This discomfort arises from holding two conflicting beliefs, values and attitudes (Harmon-Jones & Mills, 2019). As we have noted, the classical view of journalism educators is violated by AI in several respects. The journalism student was/is expected to work hard with little aid from technologies that can write for them, generate sources for them, and structure essays for them. AI technologies are doing this. It is hard for the journalism educator to embrace these technologies, or to resist them as ruining critical thinking capabilities amongst journalist trainees. Those educators who ‘surrender’ to AI find it futile to resist technologies that have become ubiquitous, while those resisting their embrace argue for a journalism training regime untampered with by such technologies. This surrender is best captured by LN4 who resignedly says “…AI is here to stay and it will keep advancing so we need to move with the times”.

There is mixed consistency in attitudes and perceptions towards AI. What we note is that AI generates mismatches amongst journalism educators. They are being introduced to a disruptive technology which violates their beliefs about journalism education. This causes discomfort, and responses to this technology are not always uniform. However, they also have no option but to respond as they have to align their own practices and perceptions of AI with external expectations at their institutions.

6. The Ethical Contestations and the Meaning of Education in the Age of AI

The discourses of AI in African contexts have been foregrounded around the language of accessibility (Ncube et al., 2025). In this regard, AI is presented as offering accessibility to often marginalised students within higher education. In other discourses, AI tools are seen as transforming higher education in several other ways, including enhancing assessment of courses, examinations and related qualifications (A. Williams, 2024). In African contexts, the discourses have almost been the same—AI can enhance higher education learning; it can transform and revolutionise higher education; and AI can bring a holistic experience for students within higher education.

In this regard, examples of AI use have been cited on many occasions. For instance, the GenAI versions have been ‘valorised’ as offering opportunities to students in terms of creative, textual production and agency. For instance, GenAIs like ChatGPT offer students an opportunity to choose their own texts and pictures to use. One Lecturer notes the following:

“We have been introduced to many AI tools that indeed have helped students. I have recommended my students to Perplexity AI, for example, which helps students discover foundational and current research in their existing fields. I encouraged my students to use this technology, and I have seen that their citations and capability to find more information have increased…but the question has been about whether this is the right way of doing education…I have been thinking if these students are paying to get machines learning”.

(Respondent LS1)

One of the most important ethical questions is the extent to which these tools are allowing students to learn. It is an ongoing debate in the African context whether AI is producing the intellectual and knowledge outcomes that higher education is expected to produce. One sceptical Lecturer in a South African university notes the following:

“AI is the opposite of what is expected in higher education. We are expected to produce solid graduates with skills. AI allows students to cheat their way through the system and allows them to graduate with nothing. These are cheating tools…and there is no other way to explain software that generates a whole essay for you without you having to spend time engaging with the literature and sharpening your arguments”.

(Respondent LS2)

But for African universities, these ethical dilemmas go beyond mere cheating. At a structural level, the question of whether these AI tools are meant for African universities is a central one. Respondents argued that AI was not made for the global south. Africa is largely a net importer of AI technologies (Arakpogun et al., 2021). These technologies are not, therefore, ‘algorithmised’ to suit African higher education and may not serve the same outcomes. One Lecturer says, “The ethical dilemma here is: who should be served by AI tools? They are not technologies for African higher education. The fact that it cannot write correctly in most of our local languages is in itself an ethical dilemma” (Respondent LS2).

The question of AI replicating algorithmic bias is central to the issue of ethics in African higher education because of the frequent absence of locally manufactured AI tools to, for example, detect plagiarism in local languages. Plagiarism cannot, for instance, be detected in African indigenous languages. Thus, at an ethical level, AI biases continue to be reproduced by AI technologies. This AI prejudice against indigenous languages and cultures is because, it was argued, AI was invented in a different cultural context: the western context. The AI bias against, or failure to comprehend, indigenous languages and cultures make its adoption in journalism training challenging. As respondent LZ3 argued, “There is a language and cultural barrier, particularly when it comes to African languages. It [AI] is not inclusive… AI misinterpret[s] other cultural contexts”. Similarly, LZ4 asserted that “…the current failure by AI to function in indigenous languages such as ChiShona or IsiNdebele…means …that AI is designed to function in English… I find AI as a Euro centered tool designed to harvest information for Eurocentric reasons”.

A ‘black market’ of AI technologies has also emerged. Small-scale AI technologies have sprouted and operate outside any formal regulation. Because they operate outside any formal structures, with, for instance, no need to register and no data provision regulations, what this means is that they also have very low threshold for data protection. They do not value privacy, and the more students and staff use them, the more they expose their data. One respondent noted the following:

“These ethical dilemmas of privacy and data protection are worse in African higher education, where students and staff use AI technologies they least understand, and where the less-protective ones are offered free online, and people simply upload their data with no guarantee of privacy” (Respondent LS3). For instance, some of the free plagiarism checkers that higher education uses, like GPTzero and Plagium, are free online. What they do not specify is how they handle metadata and other forms of data that they accumulate. Furthermore, there are ethical contestations around AI technologies that teach unethical behaviour. One Lecturer noted the following:

“We have a plagiarism detector AI, for example, a plagiarism checker. But there is also a plagiarism remover AI that helps students remove plagiarism, and for that, they do not put much effort. But the plagiarism remover is even better. There are those AI technologies that hide plagiarism. For example, Just done, Bypass, Write Human, some of which are marketed as advanced paraphrasing AI technologies. They literally teach students how to cheat systems…”.

(Respondent LS3)

The reasons for several of these unethical practices are individualistic and institutional. At an individual level, where the individual ethical threshold is low, many instances of unethical AI usage are common. At an institutional level, most African universities are yet to comprehensively respond to AI regulation, AI-driven plagiarism and any other challenges that these technologies are bringing. At an institutional level, there is also the issue of undertrained lecturers who have not been trained to deal with the gamut of AI tools their students are using. Thus, the view that AI will complement traditional teaching methods (Meyer von Wolff et al., 2020) may not be true in some African contexts. This is especially so in contexts characterised by lack of AI regulation rules, lack of lecturer training, and personal challenges that lead students to use AI and the lecturer, who is ill-equipped, to fail to recognise some of the unethical manifestations of these technologies. AI tools may set up an unethical contest between tech-savvy students and less tech-savvy lecturers. For example, the practice of AI-checked plagiarism by lecturers pits them against their own students who are using plagiarism-blocking AI technologies. Arguably, the moral and legal standing of the lecturer as an enforcer of rules will be ultimately compromised in these instances. University instructors always possess moral capital in the university systems through their knowledge and the way they share it. With AI-driven chatbots and AI writing tools on the rise, the role is equally compromised, quite severely. For instance, some of the questions one would ask are what is the role of the lecturer, when all knowledge is found online? And if all questions are answered by an AI-driven chatbot, what would be the meaning of university instruction? While these questions are asked globally, in Africa, they may take extra urgency because of the continent’s overt reliance on imported AI technologies and the absence of robust policies on AI usage within higher education systems (see Ncube et al., 2025).

7. Educators’ Knowledge Levels of AI Tools and Emerging Turf Wars

There is no denying that the journalism training landscape in Botswana, Namibia, South Africa and Zimbabwe has been reconfigured by a gamut of AI technologies (see Ncube et al., 2025). In fact, AI is significantly transforming and disrupting traditional models of journalism training in the cases under study. Despite the evident disruptions, empirical data indicates that journalism educators’ knowledge levels, institutional readiness and dexterity with these tools vary with contexts. Consequently, turf battles induced by AI pressures or external shocks are being experienced in some contexts. Therefore, discussions around journalism educators’ perceptions of AI technologies should never take for granted some of these critical factors.

The question of institutional readiness for AI integration is unclear, as discussions about AI have become common without any corresponding policies or training in place. This lack of clarity makes it challenging to assess readiness, complicating the full implementation of AI in institutions. Okolo et al. (2023) argue that effective adoption of AI requires a trained local workforce, adequate infrastructure, representative datasets, and governmental support. They further note that these critical factors are often absent in the African context, which hinders the implementation of effective AI solutions (Okolo et al., 2023). It is clear from our findings that many educational institutions are inadequately prepared for the incorporation of artificial intelligence tools, in part due to a lack of comprehensive knowledge among lecturers regarding these technologies. For example, LZ1 contends that their institution “is still very far in terms of embracing and integrating AI tools into our teaching. What needs to be done first is to have lecturers go through rigorous training on AI tools and other products so that they can be able to impart the same knowledge to students in different learning institutions”.

Most of our interviewees from Botswana, Namibia, South Africa and Zimbabwe admitted that they have limited knowledge about AI technologies and how best to incorporate them in training. These limited knowledge levels are a reflection of the fact that most journalism training institutions in Southern Africa are not yet ready for AI despite its consistent hyping. For instance, LN1 said, “I have not been trained in any of these tools”. This point was also corroborated by respondents from Zimbabwe and Botswana. For instance, LZ1 stated, “I am not using any AI tools in teaching. My understanding of AI is very limited. During my training, I never received any such kind of knowledge”. Critically, the problem of limited knowledge levels of AI is also evident in South Africa, despite the country being home to most of the leading journalism and media schools in the Southern African region. LS1 remarked the following:

“AI is a new phenomenon, and the department has not yet done anything about it. We are in a department where everyone is amused by AI. Asking questions such as do you know what AI can do? We are afraid of AI, so this means that our department is not yet ready as nothing has been done. We all realize it’s there, but they have not done anything yet”.

(interviewed 10 October 2024)

While the potential benefits of AI in education are significant, there exists a ‘cautious’ use of these tools, stemming from lecturers’ limited understanding of how these tools function. Additionally, many lecturers have expressed a need for specialised training on AI tools to facilitate their integration into pedagogical practices. This highlights the urgent necessity for educational institutions to invest in the professional development of their staff, ensuring they are well-equipped to utilise AI effectively in their teaching methodologies. Only then can we harness the full potential of these innovative tools to enhance the learning experience.

Some of the respondents from Namibia underscored that they are voluntarily attending short seminars and training courses on basic AI, particularly those offered by the Namibia Media Trust, which advocates for free expression and access to information. It is this willingness to learn which also informs the way they perceive this gamut of technologies in the learning environment. Based on these experiences, we argue that there is a need for institutions to fully empower educators so that they can competently operate in an AI environment. Some of the educators pointed out that formal training would make a huge impact and a huge difference to how AI is interpreted and incorporated in training. Due to limited institutional readiness to AI integration, some educators find the going tough in enduring and familiarising with the AI disruption. The reluctance by some journalism educators to incorporate AI in training can be a reflection of lack of adequate preparedness of some journalism training institutions in Southern Africa in exploiting opportunities provided by these technologies. This lack of grounding and understanding tends to inform how most of them perceive these technologies.

Critically, due to varying levels of awareness, understanding and engagement with the AI technologies amongst journalism and media educators in Southern Africa, this leads to creation of dual knowledge zones—those in the ‘know’ and those ‘out of the loop’. We demonstrate that the former are actively participating in adoption of AI in teaching, while the latter are still wondering and observing what AI is doing to journalism training in their respective institutions. The creation of dual or two knowledge zones contributes to the emergence of an increasingly sophisticated battle for control and retention of traditional journalism training models. In essence, there is friction between those knowledgeable of AI and those that are ‘ignorant’ of it. There is an enduring rivalry between those in the knowledgeable and ignorant zones in departments and Faculties. Some of the interviewees informed us that such rivalry manifests and openly plays out in their department or Faculty WhatsApp group platforms, and even departmental meetings. In most cases, educators use these platforms to express their fears, experiences, and aspirations about these technologies. At times, such fears about the perceived impacts of the technologies are accelerated by limited knowledge of most educators.

In most cases, those who are ignorant of AI technologies do not want to have anything to do with AI. Their assumption is that these technologies are compromising the integrity of journalism education as a profession. Some argue that the integrity of the examination system is under threat. However, we admit that while some have the knowledge, they believe that AI is here to stay and hence there is need to repurpose the technologies to the advantages of both the students and educators. Some participants from Zimbabwe and Botswana informed us that there is lack of consensus on how AI is perceived in training. There are some educators, in the Botswana case, who actually dedicate hours teaching students, especially those working on dissertations or long-term essays on how to ‘productively’ use AI. However, this makes other colleagues uncomfortable as they are of the view that such educators are encouraging students to participate in academic cheating. The lack of institutional policy on the use of these technologies accelerates the turf battles and contestations. We also note that those against these technologies are calling for abandonment of ‘take home’ assignments preferring the ‘traditional’ in-class examination model as a way of countering the AI disruption.

The dual AI knowledge zones could be partially attributed to the fact that the journalism training landscape in Southern Africa is constituted by “digital dinosaurs” and “digital natives” (see Prensky, 2001). The older generation are largely part of the digital dinosaurs and are sceptical of changes brought by AI technologies in the training field due to their habitus. Most of these educators were educated before the digital age. This is unlike the digital natives who are adept to the changing journalism training field. However, not all members of the ‘older’ generation are against AI adoption in training. Equally, not all younger educators are open to AI integration or fall in the ‘knowledgeable zone’. Some of the educators, especially from the older generation, expressed that over-reliance on AI tools can limit students’ creativity and ultimately affect the quality of storytelling. Furthermore, they are of the view that this dependence may also lead to a decline in essential journalistic skills like writing, critical thinking, and sound judgement. In fact, respondent LZ4 said “AI encourages laziness in students. At the end you risk producing ‘empty’ graduates who know nothing as AI replaced them in class”. However, there are incidences where some members of the old school are willing to learn the technologies. For example, LN2 said “No, AI is not my area of expertise, but I am open to learning and making valuable investments that will lead to mutual benefits for both students and the government”. The tensions emerging in the journalism training field affirm the argument in the domestication theory that integrating new technologies can create new burdens, tension and symmetries that necessitate further negotiations (Wang et al., 2024).

8. Discussion and Conclusions

The findings demonstrate that the invasion of the lecturers’ habitus by the external shock of AI resulted in varied responses that range from cautious acceptance to rejection. The lecturers’ responses expose intra-person tension and intra-group tension; that is, tension is evident at three levels: the individual level, the intra-group level and the intergroup level. First, at the intra-person level tension manifests at the psychological level through feelings of being overwhelmed, fear of being rendered redundant by AI, fear of decline in student critical thinking, fear of data mishandling and other unethical uses of AI such as cheating. AI thus creates psychological dissonance in educators as its affordances are a threat to the traditional role of both the educator and the learner as it threatens to replace both. AI software can write assignments and they can also mark/grade assignments. However, the greatest tension is arguably generated by the threat posed by AI to the traditional role of the lecturer. This confirms the observation that AI poses a threat to the traditional role of the instructor (see UNESCO, 2019; Chen et al., 2020).

Second, at the intra-group level, this tension manifests through the emergence of two knowledge zones that are constituted of “digital dinosaurs” and “digital natives” (see Prensky, 2001). The tension between the two zones and the intra-personal tension that manifests as psychological strain give credence to the domestication theory’s argument that integrating new technologies can create new burdens, tension and symmetries that necessitates further negotiations (Wang et al., 2024). Similarly, these tensions are a result of educators’ lack of adequate knowledge about the new AI tools which the two knowledge zones view as a threat to their profession. The rejection and partial acceptance of AI both reveal educators’ unease with AI’s invasion of their habitus in ways that challenge their traditional role and its attendant benefits. Whereas one group, ‘the digital natives’, views resisting as futile and co-option as a survival method, ‘digital dinosaurs’ instead view resistance as a survival method. One group seeks to contain AI by co-opting and adapting it to their teaching whilst the other seeks to expel it from the habitus. It is apparent that both those that resist and those that partially embrace AI are inspired by the desire to remain relevant. We also argue that the emergence of the two zones of knowledge confirms the observation that AI ushers in news forms of digital divide (see UNESCO, 2019). In addition, some lecturers’ own initiatives to learn how to adapt AI to their teaching also confirm that users are not subjects of technology. In others words, even though AI was invented in Western settings, as its opponents note, the lecturers can adapt it to their own Southern African context and to the specific context of their country and institution. In other words, whereas the respondents that rejected AI on the basis of structural issues adopted a deterministic stance, those that took the initiative to learn it and adapt it were arguably driven by a desire to be not rode roughshod by technology but by the desire to survive. Thus, they sought to adapt the technology to their contexts instead of adapting their contexts to the technology. This confirms the SST’s submission that technology is, depending on users’ political, economic, social and individual predilections, appropriated and used in different ways (Pobuda, 2022).

Third, at the intergroup level tension occurs between lecturers and students because, to suspicions of digitally enabled plagiarism, have been added suspicions of AI-generated essays and dissertations. The tension is made more palpable by fears that after using AI to write their works students then use AI again to hide that the work has been AI-generated. In other words, the fear that plagiarism- and AI-detecting software may not be helpful in detecting culprits that unethically use AI increases the tension. This threatens the traditional role of the instructor and student and is the source of the fears of cheating, fear of AI damaging students’ criticality and fears of marking machine-generated work among others. We have noted that intra-personal discomfort arises from holding two conflicting beliefs, values and attitudes (Harmon-Jones & Mills, 2019). This intra-personal discomfort inadvertently results in intra-group and intergroup tensions. For example, the classical view of journalism educators is violated by AI in several respects. The journalism student was/is expected to work hard with little aid from technologies that can write for them, generate sources for them, and structure essays for them. AI technologies are doing this. It is hard for the journalism educator to embrace this, or to resist it. Some educators view these technologies as ruining critical thinking capabilities amongst journalist trainees. Those educators who ‘adapt’ AI find it futile to resist technologies that have become ubiquitous, while those resisting their embrace argue for a journalism training regime untampered with by such technologies. However, the tensions in attitudes and perceptions towards AI are also intra-personal, as noted by one respondent who labelled the situation a ‘conundrum’. One person holds two conflicting views about AI and journalism education. For example, on the one hand they see it as a threat and on the other hand view it as a necessity. What we note is that AI generates mismatches amongst journalism educators and in individual journalism educators. They are being introduced to a disruptive technology which violates their beliefs about journalism education. This causes discomfort, and responses to this technology are not always uniform. However, they also have no option but to respond as they have to align their own practices and perceptions of AI with external expectations at their institutions.

We conclude by noting that the biggest AI-induced challenge that educators face is a psychological one—a psychological strain induced by AI’s threat to the traditional role of the instructor. The fear of being rendered redundant by technology masked as a concern for students’ criticality featured prominently in our respondents’ submissions. All the other challenges such as the absence of an AI policy, fears of unethical use of AI, lack of knowledge, etc., can arguably be traced back to this threat posed by AI to the instructor’s traditional role. The question of the role of the instructor in the AI age (or, put in other words, ‘the ethical use of AI’), we argue, should thus feature prominently in any AI policy formulation discussions. Lastly, we also conclude that this question leads us to our study’s novel observation: that the advent of AI in journalism education presents psychological strain to the educators. This is in contrast with previous studies (see UNESCO, 2019; Chen et al., 2020; Okolo et al., 2023; A. Williams, 2024) that did not consider the psychological challenge—a challenge that, as our findings show, emerges as a result of fears of being rendered redundant and fear of unethical use of AI amongst others. Future studies should consider in-depth explorations of the psychological strain that AI-induced disruptions to education cause in educators and how they manage it.