Sage Journals: Discover world-class research

Abstract

As artificial intelligence (AI) is transforming global education, it is crucial to understand the psychological barriers to its integration, given the low adoption rate among teachers. This study explores AI anxiety, as conceptualised through the Technophobia Framework, among basic school teachers in Ghana. A descriptive cross-sectional design was employed to collect data from 319 participants, with the sample size determined by an a priori power analysis. Confirmatory factor analysis (CFA) validated a four-dimensional scale measuring AI learning, job replacement, sociotechnical blindness, and configuration anxiety, demonstrating high reliability (α = .942; ω = .969). The descriptive statistics revealed moderate overall AI anxiety, while a one-way MANOVA identified a significant gender difference in sociotechnical blindness anxiety specifically, with male teachers reporting higher levels of anxiety. These findings emphasise the need for targeted professional development and gender-sensitive support strategies to address AI-related concerns and facilitate the successful integration of technology into Ghana’s educational landscape.

Plain Language Summary

Understanding Teachers’ Worries About Using Artificial Intelligence

This study looked at how basic school teachers in Ghana feel about the growing use of Artificial Intelligence (AI) in education. It focused on understanding their levels of anxiety related to learning about AI, fear of losing their jobs to AI, confusion about how AI works, and worries about setting up AI tools. Data was collected from 319 teachers, and the results showed that, overall, teachers had a moderate level of anxiety about AI. Interestingly, male teachers were more likely to feel confused or unsure about how AI fits into teaching. These findings suggest that more support and training are needed to help teachers feel confident using AI, especially in ways that consider gender differences. This research can help policymakers, school leaders, and teacher training programs make better decisions to support teachers as AI becomes more common in schools.

Keywords

AI anxiety AI learning gender job replacement anxiety sociotechnical blindness anxiety

Introduction

Artificial intelligence (AI), a transformative force in modern education, is set to revolutionise instructional delivery, enable adaptive learning pathways, and automate administrative tasks (Khairullah et al., 2025; Makinde et al., 2024; Yadav, 2025). AI is defined as the ability of computer systems to mimic human cognitive processes such as problem solving, natural language understanding, and predictive reasoning. It offers unparalleled potential to improve the quality and accessibility of education (Başer et al., 2021). AI aligns with global digital transformation trends and presents opportunities for Ghana, which has emerged as a proactive adopter of technology-integrated education. The country has made substantial investments in ICT infrastructure and educator capacity building across all academic tiers (Butakor, 2023). Notably, Ghana is one of only 15 UNESCO member states to have formally incorporated ICT into its national curriculum, reflecting a profound commitment to future-ready education (Ekumah, 2025). The national Standards-Based Curriculum introduces primary and junior high school students to foundational ICT (Quaicoe & Pata, 2018), thereby supporting SDG 4, as outlined in Target 4.4, by facilitating access to technical and other digital learning platforms.

There is a wealth of empirical evidence documenting the potential of AI to enrich learning experiences and outcomes in educational settings. Studies indicate that AI applications have a positive association with educational outcomes, significantly boosting student learning outcomes (Acquah et al., 2025; Khare et al., 2018), fostering learner-centered environments (Tuomi, 2018), and enhancing language acquisition through interactive robotics (Fryer et al., 2019). Notably, AI can transform teaching by transcending traditional pedagogical approaches (Ma & Siau, 2018). However, alongside these benefits, a critical behavioural challenge emerges: technophobia. This phenomenon, characterised by apprehension or resistance towards advanced technologies, is particularly prevalent among key educational stakeholders, such as teachers, who are often required to integrate unfamiliar digital tools into their professional practice. Technophobia introduces a new psychological barrier known as “AI anxiety,” which is the emotional unease, fear, or resistance that individuals experience when confronted with AI technologies that are perceived as complex, opaque, or threatening to human agency (Johnson & Verdicchio, 2017; Y. Y. Wang & Wang, 2022). This anxiety can substantially impede the adoption of AI and erode its potential benefits in educational settings (Bozkurt, 2023; McGrath et al., 2023).

The importance of this issue is highlighted by emerging evidence suggesting that anxiety about AI may directly affect the success of integration efforts, hindering progress towards inclusive, quality education (Kim et al., 2025). Preliminary studies, for instance, indicate that nearly 40% of educators in economies undergoing technological transition express significant apprehension regarding AI’s role in classrooms (Egara et al., 2024; Eyüp & Kayhan, 2023). It is worth noting that AI anxiety is multidimensional as contextual factors such as technological infrastructure, cultural norms, training quality, and gender dynamics can profoundly influence how it manifests (C. Chen et al., 2024; Y. M. Wang et al., 2024). For instance, in Ghana, disparities in technology access and self-efficacy could make educators more susceptible to AI-related anxieties (Butakor, 2023). However, despite its global relevance, there is a lack of empirical research in Ghana examining the mechanisms through which socio-cultural and institutional factors influence AI anxiety. This constitutes a critical research gap, particularly given the country’s ambitious agenda for ICT integration at all levels of its education system. The current study addresses this gap by investigating AI anxiety among basic school teachers in Ghana through the theoretical lens of the Technophobia Framework. This framework is particularly appropriate because it encompasses multifaceted technology-related fears and links them to adoption behaviours.

The technophobia framework provides a dynamic structure for understanding how anxieties relating to learning, job displacement, sociotechnical confusion, and configuration challenges influence professional practice collectively. Originally conceptualised by Brosnan (1998) and later refined by Rosen and Weil (1995), technophobia theory has been widely used to examine resistance to technology across various fields, providing a robust basis for the present study. In line with this, the present study, which is grounded in the Technophobia Framework, aims to explore AI anxiety among basic school teachers based on four distinct yet interrelated dimensions of AI anxiety: AI learning anxiety (apprehension about understanding and using AI tools); job replacement anxiety (fear of professional obsolescence); sociotechnical blindness anxiety (confusion regarding the interplay between AI systems and human roles); and AI configuration anxiety (discomfort related to technical setup and customization; Y. Y. Wang & Wang, 2022). Together, these dimensions offer a nuanced perspective on the psychological barriers that may hinder teachers’ adoption of AI. These four anxiety dimensions were not selected at random; rather, they represent core technophobic manifestations that are particularly relevant to educators, who must engage with AI technologies despite potential skill gaps or limited support.

Beyond theoretical contributions, literature suggests that gender stereotypes shape perceptions of AI use. Females are portrayed as having less favourable attitudes towards digital tools due to limited technology engagement (Ahn et al., 2022; Russo et al., 2025). For example, a recent study by Russo et al. (2025) revealed notable gender disparities in AI adoption, with women exhibiting greater AI anxiety, less positive attitudes towards AI, reduced AI usage, and diminished perceived AI knowledge. This underrepresentation contributes to the “technology anxiety gender gap,” which poses a significant barrier to digital inclusion and the development of 21st century skills. Of particular concern is that these gender variations may be especially prevalent in patriarchal societies in developing countries such as Ghana, where gender norms restrict access. Nevertheless, evidence-based research on gender differences in AI anxiety remains elusive in Ghana. For instance, little research has examined gender differences in AI anxiety among teachers, especially those in the basic schools. While Acquah et al. (2024) examined digital competencies and citizenship among higher education students and Ofosu-Koranteng et al. (2025) provided insights into gender variations in digital competence, neither of these studies focused on AI anxiety and particularly among teachers. As Ghana prioritises developing a knowledge-based economy and integrating ICT, it is crucial to understand gender disparities in AI anxiety among teachers.

This study aims to address this issue by investigating gender differences in AI-related anxiety among primary school teachers in Ghana. Methodologically, this study takes a rigorous quantitative approach, combining confirmatory factor analysis (CFA) with a one-way multivariate analysis of variance (MANOVA). CFA was used to validate the theoretical structure and measurement integrity of the four anxiety constructs, ensuring that each dimension was both distinct and interrelated, as predicted by the Technophobia Framework. Subsequently, one-way MANOVA was used to assess gender-based differences across these dimensions, providing insight into how sociodemographic factors may predict variations in anxiety. This dual-methods approach enhances the robustness and interpretability of the findings, enabling construct validation and group comparisons. The study aims to address the lack of research on gender differences in AI-related anxiety among primary school teachers in Ghana. Specifically, the study sought to address the following research questions:

What is the level of basic school teachers’ AI anxiety?

What are the differences in AI anxiety among basic school teachers based on their gender?

This study is of practical significance to educational policymakers, school administrators and teacher training institutions. By identifying the specific dimensions of AI anxiety that disproportionately affect educators, particularly the differences between male and female teachers, the study can inform the development of targeted professional programmes, psychosocial support interventions, and gender-sensitive strategies for integrating AI. Such insights are essential for achieving Ghana’s vision for digital education and ensuring that technological advances do not exacerbate existing inequalities. The study examines AI anxiety among basic school teachers in Ghana, exploring four key dimensions within the Technophobia Framework. It responds to the urgent need for context-specific research into barriers to technology adoption in non-Western educational settings, aiming to generate evidence that supports both theoretical advancement and practical interventions.

Literature Review

Theoretical Foundation

This study is grounded in the theoretical concept of technophobia, which is characterised by a persistent negative emotional response, including anxiety and fear, towards computer-based technologies (Brosnan, 1998; Rosen & Weil, 1995). This framework offers a psychological perspective on resistance to technological innovation, driven by perceived loss of control and low self-efficacy (Li & Huang, 2020; Y. M. Wang et al., 2024). As AI becomes more prevalent in education, technophobia principles can be used to conceptualise AI-specific anxiety among educators. This study presents AI anxiety as a modern manifestation of technophobia across four dimensions: AI Learning Anxiety, Job Replacement Anxiety, Sociotechnical Blindness Anxiety, and AI Configuration Anxiety. AI Learning Anxiety (LAX) reflects a lack of confidence in one’s ability to master complex systems (C. Chen et al., 2024; Hsu et al., 2023), whereas Job Replacement Anxiety (JRAX) represents a fear of becoming professionally obsolete (Li & Huang, 2020; Rhee & Jin, 2021). Sociotechnical blindness anxiety (STB) and AI configuration anxiety (CFAX) demonstrate technophobia’s evolution into the AI age, reflecting confusion over human-AI interaction and technical discomfort (S. Chen & Zhao, 2025; Kaya et al., 2024). Together, these dimensions form a diagnostic tool for understanding the emotional barriers to AI adoption. This study’s theoretical foundation aligns with the Technology Acceptance Model (TAM; Davis, 1989), as technophobia harms perceived ease of use and usefulness (Venkatesh & Bala, 2008; Zhou et al., 2018). This framework is particularly relevant in the context of exploring AI anxiety among teachers in Ghana’s basic education, as most are found in an environment with uneven digital competencies and varied institutional support, which can significantly hinder the integration of AI. Gender is a key analytical variable in this study’s context, as technophobia research has identified gender-specific patterns in technology adoption linked to socially constructed differences in confidence (Ayduğ & Altınpulluk, 2023; Terzi, 2020). Although findings on gender differences in AI anxiety are mixed (Banerjee & Banerjee, 2023; Falebita, 2024), the technophobia framework suggests that LAX and CFAX may be more prevalent among female teachers due to disparities in technology access. Gender provides a valuable perspective for understanding the psychological barriers to adoption, especially where gender roles influence professional development. For educational policy implications, the technophobia framework provides leaders with a diagnostic map to foster effective AI integration and enhance teacher well-being and educational outcomes. Recognising that resistance to AI stems from psychological anxiety rather than training deficits can inform more effective solutions. Findings can inform professional development programmes that address emotional barriers through mentorship schemes, phased implementation plans, and communication strategies that present AI as a pedagogical enhancer.

Empirical Review

The integration of artificial intelligence (AI) into education is now a reality, prompting a range of complex psychological responses from educators. Research on AI-related anxiety has revealed several dimensions: AI learning anxiety (apprehension about using AI tools); job replacement anxiety (fear of professional obsolescence); sociotechnical blindness anxiety (confusion about AI–human roles); and AI configuration anxiety (discomfort with technical setup; Chen & Zhao, 2025; Elfar, 2025). Studies (e.g., Ayduğ & Altınpulluk, 2023; Eyüp & Kayhan, 2023; Parviz & Arthur, 2025) have shown moderate levels of anxiety among educators, with variations observed across different fields of education and demographic groups. Eyüp and Kayhan (2023) and Ayduğ and Altınpulluk (2023) identified concerns about job security and challenges in integrating AI, while Hopcan et al. (2024) noted anxiety among pre-service teachers regarding employment prospects. Together, these findings depict the adoption of AI as a socio-psychological transition, which can be understood through the lens of technophobia, the fear of advanced technology (Brosnan, 1998; Rosen & Weil, 1995). This framework suggests that anxiety is a key barrier to the adoption of technology, affecting how educators integrate AI into their teaching methods. Gender remains an under-explored variable with conflicting evidence. While Terzi (2020) and Ayduğ and Altınpulluk (2023) found that female educators reported higher AI anxiety than males, studies by Banerjee and Banerjee (2023), Falebita (2024), and Parviz and Arthur (2025) found no significant gender differences. The different conclusions drawn from these studies may have resulted from the differences in underlying contextual factors, such as social and environmental conditions. Since most studies were conducted at the individual country level, the extent of the impacts from these facilities may vary, possibly because of context. This suggests that investigating AI anxiety while focusing on the gender variation among Ghanaian teachers, such as basic school teachers, may be more important. Notably, however, this discourse overlooks basic school teachers, who play a crucial role in the technological education of future generations. This is a particularly significant issue in Ghana and Sub-Saharan Africa.

Although studies have examined teachers’ perceptions and adoption of AI among Ghanaian teachers (Acquah et al., 2024; Akanzire et al., 2023; Ofosu-Koranteng et al., 2025; Iddrisu & Iddrisu, 2025), the specific impact of AI anxiety on basic school teachers has not been addressed. Given Ghana’s educational environment, which is characterised by resource constraints and a modernisation agenda (Baako & Abroampa, 2024; Manu et al., 2024), exploring AI anxiety constructs is imperative. AI learning anxiety can hinder professional development, while job replacement and sociotechnical anxieties can generate opposition to inclusive educational initiatives. Consequently, this study makes a decisive contribution. By moving beyond a homogenised view of AI anxiety and exploring its distinct dimensions, as well as their potential interaction with gender, among primary school teachers in Ghana, this study addresses a significant gap in the literature. The managerial implications of this study are substantial. For educational policymakers and school administrators in Ghana and similar contexts, the findings provide a nuanced diagnostic tool. Understanding whether anxiety stems from a fear of obsolescence, a lack of technical understanding, or confusion about human–AI collaboration enables targeted, effective interventions to be designed. These interventions could include gender-sensitive training programmes, clearer policy communication on the role of AI as a pedagogical aid, and improved technical support systems. These interventions could alleviate specific anxieties and foster a teaching workforce that is more confident and competent, and ready to navigate the future of education.

Methods and Materials

Sample and Study Design

This study employed a descriptive cross-sectional survey design to assess the current AI anxiety levels among basic school teachers. Following Rovai et al.'s (2014) assertion, this approach was adopted because it uses techniques and measurements that produce quantifiable values from empirical observations and measures. This design is particularly vital for critical measurement of events, subjects, objects, and ideas without manipulating the natural state of the phenomenon. The cross-sectional survey design describes a specific situation as it exists at a particular moment and necessitates direct contact with individuals whose behaviours, characteristics, and attitudes are relevant to the study (Bloomfield & Fisher, 2019; Jongbo, 2014; Rovai et al., 2014; X. Wang & Cheng, 2020). The study targeted a total population of 1,016 basic school teachers within the Cape Coast Metropolis of Ghana. A sample of 319 teachers was drawn from 33 basic schools using a simple random sampling method. This approach was adopted to guarantee that each member of the population had an equal probability of inclusion, thereby enhancing the representativeness of the sample and reducing the risk of sampling bias (Pace, 2021). The final sample size of 319 was considered statistically adequate, as it exceeded the minimum threshold of 302 participants recommended by an a priori power analysis conducted via the G*Power software. The gender distribution revealed a majority of male teachers (206, 65%), with females comprising (113, 35%) respondents.

Measure

The study utilised the AI Anxiety Scale developed by Wang and Wang (2022). The AI anxiety scale consists 21-item designed to measure four distinct dimensions of AI anxiety: “AI learning anxiety (LAX), job replacement anxiety (JRAX), sociotechnical blindness anxiety (STB), and AI configuration anxiety (CF).” Each dimension captures a specific aspect of the participants’ concerns and apprehensions regarding AI integration in education. The AI Anxiety Scale uses a 5-point Likert-type response format, ranging from “1 (strongly disagree) to 5 (strongly agree).” The refined 21-item scale, encompassing four dimensions, offered a thorough understanding of AI anxiety among basic school teachers (Y. Y. Wang & Wang, 2022). To guarantee the study’s reliability, each selected item underwent rigorous testing using “Cronbach’s alpha and McDonald’s omega coefficients,” yielding exceptional results. Notably, the scale achieved outstanding reliability scores, with Cronbach’s alpha at .942 and “McDonald’s omega” at .969. As recommended by Ravinder and Saraswathi (2020), a “McDonald’s omega coefficient” above .9 signifies a high level of data reliability. Also, “composite reliability (CR)” values for all dimensions of AI anxiety exceeded the 0.70 threshold, confirming strong internal consistency (Hair et al., 2019). In addition, convergent validity was established as “average variance extracted (AVE)” values for all constructs were above the recommended 0.50 threshold, indicating that the items adequately represented their respective constructs (Fornell & Larcker, 1981). Table 1 shows the “Cronbach alpha (α), MacDonald Omega (ω), composite reliability and convergent validity” values for each dimension of AI anxiety, providing a clear overview of the scale’s reliability.

Table 1.

Cronbach α, MacDonald ω, Composite Reliability and Convergent Validity.

S/N	Dimensions of AI anxiety	No. of items	α	ω	CR	AVE
1	AI Learning Anxiety	8	.947	.949	0.958	0.740
2	Job Replacement Anxiety	6	.907	.909	0.907	0.627
3	Sociotechnical Blindness Anxiety	4	.943	.943	0.938	0.835
4	AI Configuration Anxiety	3	.940	.942	0.941	0.843
	AI Anxiety Scale	21	.942	.969	—	—

Discriminant Validity

Discriminant validity was assessed using the “Fornell and Larcker criterion,” which requires that the square root of the AVE for each construct exceeds its correlations with other constructs (Fornell & Larcker, 1981). As shown in Table 2, the square root of the AVE for LN (0.860), JR (0.792), STB (0.914) and CF (0.918) is greater than their intercorrelation, confirming adequate discriminant validity. This indicates that each construct is empirically distinct and measures a unique aspect of AI-related anxiety, ensuring the reliability of the measurement model.

Table 2.

Discriminant Validity.

Constructs	LN	JR	STB	CF
LN	0.860
JR	0.726	0.792
STB	0.351	0.583	0.914
CF	0.439	0.576	0.897	0.918

Confirmatory Factor Analysis for AI Anxiety Scale

Confirmatory factor analysis (CFA) was performed using “JASP software version 0.19.0.0” to assess model fit (see Figure 1). Several indices were used, including “chi-squared statistic (χ²), root mean square error of approximation (RMSEA), Tucker-Lewis index (TLI), and comparative fit index (CFI).” Following the criteria established by Hu and Bentler (1999), the model was considered acceptable with CFI and TLI values greater than 0.90, “Standardised Root Mean Square Residual (SRMR)” less than 0.08 and “RMSEA” less than 0.1. The results indicated that the “AI Anxiety Scale” demonstrated a good fit to the data (χ² = 605.077, df = 71, p < .001; RMSEA = 0.06, SRMR = 0.05). In addition, the “AI Anxiety Scale” showed strong factor loadings ranging from 0.840 to 0.938, further confirming the robustness of the scale. Items that did not meet acceptable loading thresholds were systematically removed, improving the overall fit and validity of the model. This iterative refinement ensured that only good- performing items were retained, thereby improving both the precision and reliability of the measurement model.

Figure 1.

CFA model for AI anxiety scale.

Measurement Invariance

The results of the multi-group confirmatory factor analysis demonstrated evidence of gender invariance in AI Anxiety (see Table 3). The configural model exhibited an acceptable fit (CFI = 0.799, TLI = 0.742, RMSEA = 0.231, SRMR = 0.068), confirming that the overall factor structure was equivalent across male and female groups. When equality constraints were imposed on the factor loadings (metric invariance), changes in model fit indices were minimal (ΔCFI = −0.002, ΔRMSEA = −0.007), indicating that males and females interpreted the constructs similarly. Further constraining the intercepts (scalar invariance) resulted in only slight differences (ΔCFI = −0.003, ΔRMSEA = −0.002), remaining within recommended cut-off thresholds (ΔCFI ≤ −0.01, ΔRMSEA ≤ 0.015). These findings collectively suggest that the measurement of AI Anxiety is invariant across gender, allowing meaningful comparisons between male and female respondents.

Table 3.

Fit Indices for Multi-Group CFA Testing Gender Invariance in AI Anxiety.

Invariance level	CFI	ΔCFI	TLI	ΔTLI	SRMR	ΔSRMR	RMSEA	ΔRMSEA	AIC	ΔAIC	χ² (Δχ²)	df	Δdf
Configural invariance	0.799	—	0.742	—	0.068	—	0.231	—	8,848.903	—	1,355.551	142	—
Metric invariance	0.797	−0.002	0.757	0.015	0.071	0.003	0.224	−0.007	8,847.137	−1.766	1,373.784 (18.23***)	152	10
Scalar invariance	0.794	−0.003	0.754	−0.003	0.073	0.002	0.222	−0.002	8,849.504	2.367	1,394.263 (20.48***)	162	10
Cut-off Criteria	≥0.95	Δ ≤ −0.01	≥0.95	Δ ≤ 0.01	≤0.08	Δ ≤ 0.030	≤0.06	Δ ≤ 0.015	—	—

Note. Δ = Change relative to the previous model. ML estimator used. CFI = Comparative Fit Index; TLI = Tucker–Lewis Index; SRMR = Standardised Root Mean Square Residual; RMSEA = Root Mean Square Error of Approximation; AIC = Akaike Information Criterion; χ² = Chi-Square Statistic.

***

p < .001.

Data Collection Procedure

Before embarking on the study, the researchers thoroughly understood the principles of research ethics and the specifics of the survey instrument. They personally engaged with basic school teachers across various regions of Ghana, ensuring effective data collection, and selected a representative sample. The researchers, using an ethical approach, administered the questionnaires, allowing each respondent an adequate 30 to 40 min to thoughtfully complete the survey. Once the questionnaires were returned, the researchers meticulously reviewed each one, verifying accuracy and completeness to ensure the data were primed for analysis.

Data Analysis Strategy

The data gathered was processed and enhanced using the “SPSS software (version 26).” Initially, the demographic characteristics of the participants were subjected to analysis utilising frequencies and percentages. Data gathered on the first objective, which aimed at examining basic school teachers’ level of AI anxiety, was analysed through the application of descriptive statistics, involving means and standard deviations. Furthermore, to address the second research objective, which sought to determine the differences in AI anxiety among basic school teachers, a “one-way MANOVA” was conducted. Despite the gender imbalance, a “one-way MANOVA” remains appropriate for examining differences in AI anxiety between male and female teachers, as it accommodates unequal group sizes (Pallant, 2020).

Ethical Consideration

It is essential to emphasise the significance of maintaining mutual respect within the framework of our study. “Institutional Review Board (IRB) of the researchers’ university (Ethical Clearance—ID [“UCCIRB/CES/2023/169”]).” All respondents were willingly recruited to contribute data and were fully briefed on the study, with the freedom to withdraw at any stage. Also, respondents were made to sign a written informed consent form. Privacy protection was paramount, and comprehensive measures were implemented to ensure the safeguarding of all data and identities. The researchers were unwavering in their commitment to preventing any form of harm and were dedicated to steering clear of any factors that could potentially lead to discomfort throughout the study. Our devotion to honesty and transparency has been the cornerstone of our methodology, from the clear and open approach in our methods to genuine and forthright reporting, aligning with both institutional and legal standards to uphold the integrity of our research.

Results

Basic School Teachers’ Level of AI Anxiety

This objective examined basic school teachers’ level of AI anxiety. Prior to that, the normality of the data was checked to ascertain whether it deviated from a normal distribution. Table 4 shows the results for skewness and kurtosis, as well as the mean and standard deviation for the dimensions of AI anxiety.

Table 4.

Descriptive Statistics.

Constructs	N	Minimum	Maximum	Mean	SD	Skewness		Kurtosis
Constructs	Statistic	Statistic	Statistic	Statistic	Statistic	Statistic	Std. error	Statistic	Std. error
LAX	319	1.00	5.00	3.53	1.02	−0.962	0.137	−0.419	0.272
JRAX	319	1.00	5.00	3.51	1.07	−0.870	0.137	−0.614	0.272
STB	319	1.00	5.00	3.44	1.02	−0.851	0.137	−0.734	0.272
CF	319	1.00	5.00	3.41	1.02	−0.913	0.137	−0.520	0.272
Overall, AI Anxiety				3.47	1.03

Note. Scale: 1.00–1.49 (Very Low); 1.50–2.49 (Low); 2.50–3.49 (Moderate); 3.50–4.49 (High); 4.50–5.00 (Very High). SD = Standard deviation.

Source. Arkorful et al. (2025).

The normality of the distribution of AI anxiety levels among basic school teachers reveals important characteristics in Table 4. The skewness values for all measured categories (“Learning Anxiety—LAX, Job Replacement Anxiety—JRAX, Sociotechnical Blindness Anxiety—STB, and AI Configuration Anxiety—CF”) are negative, indicating a leftward skew in the distribution. Specifically, these values range from −0.870 to −0.962, suggesting that a greater number of teachers reported higher anxiety levels, with fewer indicating very low anxiety. Additionally, the negative kurtosis values indicate a flatter distribution compared to a normal distribution. While the data shows a tendency toward higher anxiety levels, it does not deviate significantly from normality, allowing for valid statistical analysis.

The descriptive statistics provide a comprehensive overview of AI anxiety levels among the respondents. The sample size (N) for each anxiety category is 319, and the mean anxiety levels range from 3.41 (CF) to 3.53 (LAX). All categories fall within the “Moderate” range (2.50–3.49) to “High” range (3.50–4.49) according to the defined scale. Notably, the highest mean score of 3.53 for LAX indicates a significant level of learning anxiety related to AI. This is closely followed by JRAX and STB, which also reflect high levels of anxiety regarding job replacement and sociotechnical blindness, respectively. In all, the basic school teachers had a moderate level of AI anxiety.

Differences in AI Anxiety Among Basic School Teachers Based on Gender

This objective sought to determine the differences in AI anxiety among basic school teachers based on gender. Preliminary analyses were conducted to ensure that the assumptions of the “one-way MANOVA” were met. Multicollinearity was assessed by checking the correlation between the dimensions of AI anxiety. According to Anderson et al. (2011), multicollinearity is a concern when the correlation between any two variables exceeds 0.90. The results indicated that none of the AI anxiety dimensions were highly correlated, suggesting the absence of multicollinearity (see Table 5).

Table 5.

Pearson Correlations Among the AI Anxiety Dimensions (N = 319).

Variable	1. LN	2. JR	3. STB	4. CF
1. LN	—
2. JR	.660**	—
3. STB	.378**	.547**	—
4. CF	.387**	.545**	.867**	—

Correlation is significant at the 0.01 level (2-tailed).

In addition, descriptive statistics were computed to explore gender-based differences in AI anxiety. This provided a general overview of the distribution of scores across the different dimensions for male and female teachers (see Table 6). Assumption tests were also performed prior to the main “one-way MANOVA.” Levene’s test for the equality of error variances revealed no violation, indicating that the assumption of “homogeneity of variances” was met. However, “Box’s Test of Equality of Covariance Matrices” was significant (Box’s M = 53.144, F[10, 252,204.372] = 5.233, p < .001), suggesting a violation of the assumption of equal covariance matrices across gender groups. Due to this violation, “Pillai’s Trace” was employed for interpreting the “one-way MANOVA” results, as it is considered robust in the presence of “unequal covariance matrices” (Pallant, 2020).

Table 6.

Descriptive Statistics for the Difference in AI Anxiety Based on Gender.

Constructs	Gender	Mean	SD	N
LN	Male	3.58	0.96	206
	Female	3.44	1.12	113
	Total	3.53	1.02	319
JR	Male	3.59	0.99	206
	Female	3.36	1.19	113
	Total	3.51	1.07	319
STB	Male	3.56	0.98	206
	Female	3.22	1.07	113
	Total	3.44	1.02	319
CF	Male	3.47	0.97	206
	Female	3.30	1.09	113
	Total	3.41	1.02	319

Note. Scale: 1.00−1.49 (Very Low); 1.50−2.49 (Low); 2.50−3.49 (Moderate); 3.50−4.49 (High); 4.50−5.00 (Very High).

The descriptive statistics for AI anxiety by gender, as shown in Table 6, demonstrate that male teachers exhibit higher anxiety levels than their female counterparts in all assessed categories. For Learning Anxiety, male teachers reported a mean level of 3.58 (SD = 0.96), categorising their anxiety as high. In contrast, female teachers had a mean of 3.44 (SD = 1.12), also falling within the high range. This trend continues with Job Replacement Anxiety, where males scored 3.59 (SD = 0.99), again indicating as a high anxiety, compared to 3.36 (SD = 1.19) for females, which is categorised as moderate. The most pronounced difference appears in Sociotechnical Blindness Anxiety, where male teachers reported a mean of 3.56 (SD = 0.97970), indicating high anxiety, versus 3.22 (SD = 1.07) for females, categorised as moderate. Similarly, in AI Configuration Anxiety, male teachers scored 3.47 (SD = 0.97), also in the high range, while female teachers had a mean of 3.30 (SD = 1.09), categorising their anxiety as moderate. The findings indicate a gender disparity in basic school teachers’ anxiety about AI, underscoring the necessity of a one-way MANOVA to investigate the statistically significant variation in AI literacy according to gender. Table 7 displays the MANOVA results.

Table 7.

Multivariate Test.

Effect	Value	F	Hypothesis df	Error df	Sig.	Partial eta squared
Gender
Pillai’s Trace	0.039	3.165	4.000	314.000	0.014	.039
Wilks’ Lambda	0.961	3.165	4.000	314.000	0.014	.039
Hotelling’s Trace	0.040	3.165	4.000	314.000	0.014	.039
Roy’s Largest Root	0.040	3.165	4.000	314.000	0.014	.039

The Pillai’s Trace result showed that there are significant differences in AI anxiety based on gender (p = .014, η² = .039; see Table 7). All of these tests yielded significant values (p < .05) with partial eta squared values of .039, indicating a small effect size. In particular, Table 8 demonstrates significant gender differences between-subject impacts across a range of AI anxiety variables. For “Learning Anxiety (LN),” the F value was 1.405 (p = .237, η² = .004), indicating no significant effect. Job Replacement Anxiety (JR) had an F value of 3.391 (p = .067, η² = .011), which approaches significance but does not meet the adjusted criterion. In contrast, “Sociotechnical Blindness Anxiety (STB)” exhibited a significant F value of 8.040 (p = .005, η² = .025). AI “Configuration Anxiety (CF)” showed an F value of 2.079 (p = .150, η² = .007), indicating no significant effect. These results suggest that while gender influences AI anxiety levels, a statistically significant difference was observed only in “Sociotechnical Blindness Anxiety” (see Table 8).

Table 8.

Tests of Between-Subjects Effects.

Source	Dependent variable	Type III sum of squares	df	Mean square	F	Sig.	Partial eta squared
Corrected Model	LN	1.468	1	1.468	1.405	0.237	.004
	JR	3.875	1	3.875	3.391	0.067	.011
	STB	8.223	1	8.223	8.040	0.005	.025
	CF	2.142	1	2.142	2.079	0.150	.007
Intercept	LN	3,599.391	1	3,599.391	3,445.591	<0.001	.916
	JR	3,518.951	1	3,518.951	3,078.887	<0.001	.907
	STB	3,358.414	1	3,358.414	3,283.897	<0.001	.912
	CF	3,341.734	1	3,341.734	3,244.530	<0.001	.911
Gender	LN	1.468	1	1.468	1.405	0.237	.004
	JR	3.875	1	3.875	3.391	0.067	.011
	STB	8.223	1	8.223	8.040	0.005	.025
	CF	2.142	1	2.142	2.079	0.150	.007
Error	LN	331.150	317	1.045
	JR	362.309	317	1.143
	STB	324.193	317	1.023
	CF	326.497	317	1.030
Total	LN	4,312.800	319
	JR	4,286.778	319
	STB	4,109.444	319
	CF	4,034.889	319
Corrected Total	LN	332.617	318
	JR	366.184	318
	STB	332.416	318
	CF	328.639	318

Note. Bonferroni adjustment: p < .0125 (0.05/4 = 0.0125).

Discussion

This study critically explores AI anxiety among basic school teachers in Ghana, employing a robust methodological framework based on confirmatory factor analysis (CFA) and one-way MANOVA. The findings shed light on the complex emotional landscape that shapes educators’ readiness for AI integration, offering nuanced insights that both align with and challenge existing literature on this topic. The discussion interprets these results through the theoretical lens of technophobia and situates them within the Ghanaian sociocultural context, elucidating their significant implications for policy and practice. CFA confirmed the validity and reliability of the four-dimensional AI anxiety scale, with all factor loadings exceeding 0.84, indicating a robust measurement model. Descriptive findings revealed a moderate to high level of overall AI anxiety among Ghanaian basic school teachers. This collective apprehension is a quintessential manifestation of technophobia: a persistent negative emotional response to advanced technology, driven by perceived low self-efficacy and a potential loss of control (Brosnan, 1998; Li & Huang, 2020). Delving into the specific dimensions, AI Learning Anxiety (LAX) emerged with the highest mean score. This signifies a pronounced apprehension about understanding and utilising AI tools, which directly reflects the low self-efficacy component of the technophobia framework. Teachers feel inadequately prepared to master these complex systems, a concern that is exacerbated in contexts such as Ghana, where professional development on digital tools is inconsistent and under-resourced (Manu et al., 2024).

Similarly, Job Replacement Anxiety (JRAX) was notably high, echoing a fear of professional obsolescence that aligns with global concerns identified by Hopcan et al. (2024) and Eyüp and Kayhan (2023). This dimension of technophobia is rooted in the perceived threat AI poses to one’s professional identity and job security. The anxieties concerning Sociotechnical Blindness (STB) and AI Configuration, while slightly lower, remain in the high range. These dimensions represent the evolution of technophobia into the AI age, capturing confusion about the human-AI interplay and discomfort with the technical mechanics of AI systems (S. Chen & Zhao, 2025; Kaya et al., 2024). The strong correlation between STB and CF is particularly telling; confusion about the societal role of AI is intrinsically linked to anxiety about its technical setup. This suggests interventions must address both conceptual understanding and hands-on skills simultaneously. A pivotal finding of this study is the nuanced role of gender, but this difference was statistically significant only for STB, confirming significant gender differences exclusively for sociotechnical blindness anxiety, which adds a critical layer to the understanding of AI anxiety. This study found that male teachers reported higher mean scores across all four anxiety dimensions is contrary to previous studies that found higher general AI anxiety among females (Ayduğ & Altınpulluk, 2023; Terzi, 2020). This intriguing discrepancy demands a context-specific explanation rooted in sociocultural norms and role expectations within Ghanaian society.

The finding that Ghanaian male basic school teachers experienced greater anxiety regarding STB is contrary to Ayduğ and Altınpulluk (2023), and this may be linked to culturally constructed gender roles. In many contexts, including Ghana, technological proficiency is often considered a masculine trait. This creates unique pressures for male teachers. When confronted with AI, the implications of which are vast and unclear, male educators may experience heightened anxiety due to the implicit pressure to master it despite being unable to do so. The resulting conflict between societal expectations and personal understanding can manifest as sociotechnical blindness and anxiety. Female teachers, who are less burdened by the pressure to be technological experts, may therefore approach AI with less anxiety about its implications, resulting in lower STB. Notably, there are no gender differences in AI learning, job replacement, and configuration anxieties, suggesting that these fears are universal concerns for all teachers, transcending gender. This aligns with the findings of Banerjee and Banerjee (2023) and Falebita (2024), who suggest that core technophobic responses to AI may be more universal than gender-specific in professional settings.

These findings are consistent with international research indicating moderate to high levels of anxiety about AI among educators (Ayduğ & Altınpulluk, 2023; Eyüp & Kayhan, 2023). Job Replacement Anxiety corroborates the concerns about professional obsolescence raised by Hopcan et al. (2024). However, this study challenges the idea that women report higher levels of AI anxiety, suggesting that anxiety can be gendered in nature, while overall levels remain similar. The significant difference in STB, but not in other dimensions, suggests that sociocultural factors influence how anxiety is expressed rather than its intensity. This helps to explain discrepancies in previous literature by highlighting how contextual factors, including gender norms and educational technology infrastructure, influence technophobia. This study reveals the nuanced impact of gender on AI anxiety dimensions within a specific cultural context, confirming technophobia as a useful framework for understanding barriers to adoption. Addressing these barriers requires strategies that are both technically sound and culturally sensitive to support teachers in navigating educational futures.

Theoretical and Methodological Contributions

This study makes pivotal contributions to the understanding of AI anxiety within educational technology and technophobia in the age of AI. Through confirmatory factor analysis, it validates the multidimensional structure of AI anxiety, confirming a four-factor model comprising AI Learning Anxiety (LAX), Job Replacement Anxiety (JRAX), Sociotechnical Blindness Anxiety (STB), and AI Configuration Anxiety (CFAX). This model can be used as a diagnostic tool to identify emotional barriers among educators. The scale’s robust psychometric properties and discriminant validity offer researchers a dependable instrument for evaluation across cultural contexts (Fornell & Larcker, 1981). This study extends technophobia theory in Sub-Saharan African education, with findings from Ghanaian primary school teachers challenging Western-centric assumptions. Contrary to narratives associating technology apprehension with females, male teachers show higher Sociotechnical Blindness Anxiety (Ayduğ & Altınpulluk, 2023; Terzi, 2020). In the Ghanaian cultural context, masculine expectations of technological mastery create a psychological burden when confronting AI. This necessitates a theoretical revision that considers technophobia to be dynamically mediated by local gender norms (Banerjee & Banerjee, 2023; Falebita, 2024).

Methodologically, this study provides a sophisticated analytical approach by applying MANOVA to address AI anxiety variation between male and female basic school teachers, showing that gender differences are dimension-specific (Pallant, 2020). This approach clarifies the phenomenological nature of anxiety and provides a model for understanding variations in gender subgroups within psychological constructs. The study presents higher AI anxiety among males compared to their female counterparts, and this presents a psychological response which could stem from perceived low self-efficacy and professional identity threat (Li & Huang, 2020; Y. M. Wang et al., 2024). This therefore shows that AI integration must address emotional and cognitive barriers through gender-specific interventions such as mentorship schemes that reframe AI as a pedagogical enhancer. Overall, by linking technophobia with AI-related challenges, the study provides a theoretical framework for understanding educator resistance and for developing effective implementation strategies.

Practical Implications

This study provides a critical, context-specific contribution to the burgeoning field of AI in education by validating a multidimensional model of AI anxiety within Ghana’s basic education sector. Moving beyond Western-centric literature, our findings reveal that teacher readiness depends on emotional, technical, and sociocultural factors, rather than just access to technology. The validated four-dimensional anxiety scale, which covers anxieties relating to AI learning, job replacement, sociotechnical blindness, and configuration, provides educators and policymakers with a diagnostic tool with which to target interventions. This highlights the need for a culturally attuned strategy to foster sustainable AI adoption.

For national policymakers at the Ministry of Education (MoE) and the Ghana Education Service (GES), these findings are a clarion call to action. The widespread anxiety requires the development of a national AI literacy framework for teachers. This policy should present AI as a pedagogical collaborator to mitigate anxiety about job replacement. Our analysis revealed a gender disparity in sociotechnical anxiety, with male teachers showing greater confusion about human–AI interaction. This finding, which is rooted in cultural pressures for technological mastery, suggests that a one-size-fits-all approach is ineffective. Policies must therefore mandate gender-sensitive professional development modules that address these sociotechnical concerns through scenario-based learning and dialogue about the role of AI.

For educational managers and school leadership, it is imperative to translate national policy into practical support systems. The most effective way to address anxieties surrounding AI learning is to establish robust Professional Learning Communities (PLCs). Head teachers must champion PLCs as spaces for peer mentoring, demonstrations of AI tools, and collaborative problem solving. This approach fosters self-efficacy, countering technophobia (Brosnan, 1998), and is consistent with leadership. Leaders must create an environment that encourages experimentation and addresses concerns, reframing AI as a tool that enhances professionalism.

For non-governmental organisations and teacher training institutions, this study highlights the importance of partnerships. Universities must redesign their curricula to incorporate AI pedagogy and ethics, ensuring that new teachers enter the profession ready to confidently embrace AI. NGOs and EdTech partners can support the development of contextually appropriate AI training resources for use in Ghanaian classrooms. These partnerships are vital for ensuring equitable access to training programmes. The key beneficiaries are Ghanaian primary school pupils. A teaching force that is confident and critically AI-literate will create more engaging and effective learning environments. Addressing anxiety barriers paves the way for technology to enhance human connection and pedagogical excellence.

Finally, both the Ministry of Education and GES should establish mechanisms to monitor and evaluate the implementation of AI-related teacher support initiatives. This includes developing metrics to assess reductions in AI anxiety over time and determining the effectiveness of interventions at both the regional and national levels. Data from such evaluations can guide iterative improvements in policy and practice, ensuring that support systems evolve with teachers’ changing needs and the rapid pace of technological advancement.

Limitations of the Study and Suggestions for Future Research

This study validated a multidimensional AI anxiety scale, confirming the existence of four dimensions and providing a reliable, cross-cultural diagnostic tool. Confirmatory factor analysis and MANOVA revealed dimension-specific variations, including higher anxiety about sociotechnical blindness among male teachers. This offers insights for AI integration interventions in Ghana’s education system. However, the cross-sectional design limits the ability to draw causal inferences and understand how teachers’ anxieties evolve. Also, focusing on Ghanaian basic school teachers restricts the generalisability of the findings, while the self-reported data may introduce social desirability bias. Given the limited examination of demographic variables and anxiety relationships in this study, future research should employ longitudinal designs and diverse methods to improve our understanding of AI anxiety and develop support frameworks.

Conclusions

This study sheds light on the challenges of technology integration in developing countries by examining AI anxiety among Ghanaian basic school teachers. The findings revealed moderate levels of AI anxiety overall, with specific concerns relating to AI-based learning and job replacement. This highlights the importance of incorporating interventions into teacher training programmes to enhance AI literacy and address concerns about job security. Male teachers exhibited higher socio-technical blindness anxiety than females, challenging existing literature and suggesting that the relationship between gender and AI anxiety is complex and shaped by unique Ghanaian factors. The absence of gender differences in other AI anxiety dimensions is consistent with some studies but inconsistent with others, emphasising the need for localised research to inform the integration of AI in education.

Footnotes

ORCID iDs

Francis Arthur

Eric Boateng

Ernest Opoku

Francis Obeng Gyedu

Stanley Asare-Bediako

Emmanuel Quayson

Iddrisu Salifu

Emmanuel Rungson Attom

Sharon Abam Nortey

Ethical Considerations

Ethical approval was sought from the “Institutional Review Board (IRB) of the University of Cape Coast (Ethical Clearance—ID [UCCIRB/CES/2023/169]).”

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data will be made available upon request from the corresponding author.

References

Acquah

B. Y. S.

Arthur

Salifu

Mensah

Opoku

Nortey

S. A.

Tetteh

S. A.

(2025). Modelling economics students' use of ChatGPT and academic performance: Insights from self-determination theory and epistemic curiosity. Discover Artificial Intelligence, 5(1), 1–24. https://doi.org/10.1007/s44163-025-00373-y

Acquah

B. Y. S.

Arthur

Salifu

Quayson

Nortey

S. A.

(2024). Preservice teachers' behavioural intention to use artificial intelligence in lesson planning: A dual-staged PLS-SEM-ANN approach. Computers and Education: Artificial Intelligence, 7, 100307. https://doi.org/10.1016/j.caeai.2024.100307

Ahn

Kim

Sung

(2022). The effect of gender stereotypes on artificial intelligence recommendations. Journal of Business Research, 141, 50–59. https://doi.org/10.1016/j.jbusres.2021.12.007

Akanzire

B. N.

Nyaaba

Nabang

(2023). Perceptions and preparedness: Exploring teacher educators’ views on integrating generative AI in colleges of education Ghana. Journal of Digital Educational Technology, 5(1), ep2508.

Anderson

D. R.

Sweeney

D. J.

Williams

T. A.

(2011). Statistics for business and economics. South-Western, Cengage Learning.

Arkorful

Arthur

Boateng

Ofosu-Koranteng

Salifu

Attom

E. R.

Tetteh

S. A.

Quayson

Asare-Bediako

Nortey

S. A.

(2025). Exploring artificial intelligence literacy among basic school teachers in Ghana. Discover Education, 4(1), 250. https://doi.org/10.1007/s44217-025-00630-3

Ayduğ

Altınpulluk

(2023). Investigation of pre-service teachers’ artificial intelligence anxiety levels [Conference session]. The 36th Asian Association of Open Universities Annual Conference- AAOU 2023, İstanbul. https://www.researchgate.net/publication/379154215

Baako

Abroampa

W. K.

(2024). Context matters: Exploring teacher and learner contexts in ICT integration in slum public basic schools in Ghana. Cogent Education, 11(1), Article 2342637. https://doi.org/10.1080/2331186X.2024.2342637

Banerjee

(2023). College teachers’ anxiety towards artificial intelligence: A comparative study. Research Review International Journal of Multidisciplinary, 8(5), 36–43. https://doi.org/10.31305/rrijm.2023.v08.n05.005

10.

Başer

Altuntaş

S. B.

Kolcu

Özceylan

(2021). Artificial intelligence anxiety of family physicians in Turkey. Progress in Nutrition, 23(Suppl 2), Article e2021275. https://doi.org/10.23751/pn.v23iS2.12003

11.

Bloomfield

Fisher

M. J.

(2019). Quantitative research design. Journal of the Australasian Rehabilitation Nurses Association, 22(2), 27–30.

12.

Bozkurt

(2023). ChatGPT, üretken yapay zeka ve algoritmik paradigma değişikliği [ChatGPT, Generative Artificial Intelligence, and the Algorithmic Paradigm Shift]. Alanyazın, 4(1), 63–72. https://doi.org/10.59320/alanyazin.1283282

13.

Brosnan

M. J.

(1998). The impact of computer anxiety and self-efficacy upon performance. Journal of Computer Assisted Learning, 14(3), 223–234. https://doi.org/10.1046/j.1365-2729.1998.143059.x

14.

Butakor

P., K.

(2023). Exploring pre-service teachers’ beliefs about the role of artificial intelligence in higher education in Ghana. International Journal of Innovative Technologies in Social Science, 3(39). https://doi.org/10.31435/rsglobal_ijitss/30092023/8057

15.

Chen

Wei

(2024). From anxiety to action: Exploring the impact of artificial intelligence anxiety and artificial intelligence self-efficacy on motivated learning of undergraduate students. Interactive Learning Environments, 41(13), 8438–8446. https://doi.org/10.1080/10494820.2024.2440877

16.

Chen

Zhao

(2025). Explore the AI anxiety attitude of E-commerce employees. International Journal of Human–Computer Interaction, 41(13), 8438–8446. https://doi.org/10.1080/10447318.2024.2409470

17.

Davis

F. D.

(1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. https://doi.org/10.1016/0737-2301(89)10001-2

18.

Egara

F. O.

Mosimege

(2024). Exploring the integration of artificial intelligence-based ChatGPT into mathematics instruction: Perceptions, challenges, and implications for educators. Education Sciences, 14(7), Article 742. https://doi.org/10.3390/educsci14070742

19.

Ekumah

R. C.

(2025). An analysis of Ghana’s national ICT policy in education [Unpublished Doctoral dissertation, American University of Beirut].

20.

Elfar

E. E.

(2025). Exploring the impact of AI awareness on AI anxiety: The moderating role of perceived organizational support. Management & Sustainability: An Arab Review. Advance online publication. https://doi.org/10.1108/MSAR-01-2025-0008

21.

Eyüp

Kayhan

(2023). Pre-service Turkish language teachers’ anxiety and attitudes toward artificial intelligence. International Journal of Education and Literacy Studies, 11(4), 43–56. https://doi.org/10.7575/aiac.ijels.v.11n.4p.43

22.

Falebita

O. S.

(2024). Assessing the relationship between anxiety and the adoption of Artificial Intelligence tools among mathematics preservice teachers. Interdisciplinary Journal of Education Research, 6, 1–13. https://doi.org/10.38140/ijer-2024.vol6.20

23.

Fornell

Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50.

24.

Fryer

L. K.

Nakao

Thompson

(2019). Chatbot learning partners: Connecting learning experiences, interests and competence. Computers in Human Behaviors, 93, 279–289. https://doi.org/10.1016/j.chb.2018.12.023

25.

Hair

J. F.

Risher

J. J.

Sarstedt

Ringle

C. M.

(2019). When to use and how to report the results of PLS-SEM. European Business Review, 31(1), 2–24.

26.

Hopcan

Türkmen

Polat

(2024). Exploring the artificial intelligence anxiety and machine learning attitudes of teacher candidates. Education and Information Technologies, 29(6), 7281–7301. https://doi.org/10.1007/s10639-023-12086-9

27.

Hsu

T. C.

Hsu

T. P.

Lin

Y. T.

(2023). The artificial intelligence learning anxiety and self-efficacy of in-service teachers taking AI training courses [Conference session]. 2023 International Conference on Artificial Intelligence and Education (ICAIE) (pp. 97–101). IEEE.

28.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

29.

Iddrisu

H. M.

Iddrisu

S. A.

(2025). Teacher readiness to use AI tools in the classroom. International Journal of Research in Education and Science, 11(3), 494–510. https://doi.org/10.46328/ijres.1314

30.

Johnson

D. G.

Verdicchio

(2017). AI anxiety. Journal of the Association for Information Science and Technology, 68(9), 2267–2270. http://doi.org/10.1002/asi.23867

31.

Jongbo

O. C.

(2014). The role of research design in a purpose driven enquiry. Review of Public Administration and Management, 400(3615), 1–8.

32.

Kaya

Aydin

Schepman

Rodway

Yetişensoy

Demir Kaya

(2024). The roles of personality traits, AI anxiety, and demographic factors in attitudes toward artificial intelligence. International Journal of Human–Computer Interaction, 40(2), 497–514.

33.

Khairullah

S. A.

Harris

Hadi

H. J.

Sandhu

R. A.

Ahmad

Alshara

M. A.

(2025). Implementing artificial intelligence in academic and administrative processes through responsible strategic leadership in the higher education institutions. Frontiers in Education, 10, Article 1548104.

34.

Khare

Stewart

Khare

(2018). Artificial intelligence and the student experience: An institutional perspective. The International Academic Forum (IAFOR). https://doi.org/0.22492/ije.6.3.04

35.

Kim

J. J.

Soh

Kadkol

Solomon

Yeh

Srivatsa

A. V.

Nahass

G. R.

Choi

J. Y.

Lee

Nyugen

Ajilore

(2025). AI anxiety: A comprehensive analysis of psychological factors and interventions. AI and Ethics, 5, 3993–4009. https://doi.org/10.1007/s43681-025-00686-9

36.

Huang

J. S.

(2020). Dimensions of artificial intelligence anxiety based on the integrated fear acquisition theory. Technology in Society, 63, Article 101410.

37.

Makinde

A. I.

Adeleye

S. A.

Oronti

A. O.

Jimoh

I. T.

(2024). Revolutionizing education: AI in next-generation mobile management. In Sur

S. N.

Imoize

A. L.

Bhattacharya

Kandar

Banerjee

J. S.

(Eds.), Artificial intelligence for wireless communication systems: Technology and applications (pp. 103–125). CRC Press.

38.

Manu

Ampomah

Akyina

K. O.

Antwi

(2024). Education and technology in Ghana: Understanding the centrality of technology integration in the classroom and beyond. American Journal of Educational Research, 12(10), 402–412. https://doi.org/10.12691/education-12-10-2

39.

Siau

K. L.

(2018, May 17–18). Artificial intelligence impacts on higher education [Conference session]. Thirteenth Annual Midwest Association for Information Systems Conference, St. Louis, Missouri, USA.

40.

McGrath

Pargman

T. C.

Juth

Palmgren

P. J.

(2023). University teachers’ perceptions of responsibility and artificial intelligence in higher education-an experimental philosophical study. Computers and Education: Artificial Intelligence, 4, Article 100139. https://doi.org/10.1016/j.caeai.2023.100139

41.

Ofosu-Koranteng

Arthur

Salifu

Quayson

Boateng

Arkorful

Asare-Bediako

Attom

R. E.

(2025). Gender-driven variations in digital competence among economics students in a Ghanaian higher education institution. Educational Media International, 62(3), 267–288. https://doi.org/10.1080/09523987.2025.2506131

42.

Pace

D. S.

(2021). Probability and non-probability sampling-an entry point for undergraduate researchers. International Journal of Quantitative and Qualitative Research Methods, 9(2), 1–15.

43.

Pallant

(2020). SPSS survival manual: A step by step guide to data analysis using IBM SPSS (7th ed.). Routledge. https://doi.org/10.4324/9781003117452

44.

Parviz

Arthur

(2025). AI anxiety in english language education: A study of Iranian EFL teachers' perceptions and demographic influences. International Journal of Computer-Assisted Language Learning and Teaching (IJCALLT), 15(1), 1–21.

45.

Quaicoe

J. S.

Pata

(2018). Basic school teachers’ perspective to digital teaching and learning in Ghana. Education and Information Technologies, 23(3), 1159–1173.https://doi.org/10.1007/s10639-017-9660-8

46.

Ravinder

E. B.

Saraswathi

A. B.

(2020). Literature review of Cronbach alpha coefficient (Ω) and Mcdonald’s omega coefficient (Ω). European Journal of Molecular & Clinical Medicine, 7(6), 2943–2949.

47.

Rhee

Jin

(2021). The Effect of job anxiety of replacement by artificial intelligence on organizational members’ job satisfaction in the 4th industrial revolution era: The moderating effect of job uncertainty. Journal of Digital Convergence, 19(7), 1–9.

48.

Rosen

L. D.

Weil

M. M.

(1995). Computer availability, computer experience and technophobia among public school teachers. Computers in Human Behavior, 11(1), 9–31. https://doi.org/10.1016/0747-5632(94)00018-D

49.

Rovai

A. P.

Baker

J. D.

Ponton

M. K.

(2014). Social science research design and statistics. Watertree Press LLC.

50.

Russo

Romano

Clemente

Iacovone

Gladwin

T. E.

Panno

(2025). Gender differences in artificial intelligence: The role of artificial intelligence anxiety. Frontiers in Psychology, 16, Article 1559457. https://doi.org/10.3389/fpsyg.2025.1559457

51.

Terzi

(2020). An adaptation of artificial intelligence anxiety scale into Turkish: Reliability and validity study. International Online Journal of Education and Teaching, 7(4), 1501–1515.

52.

Tuomi

(2018). The impact of artificial intelligence on learning, teaching, and education. Publications Office of the European Union. http://mx.nthu.edu.tw/∼cshwang/data-economics/courseinfoecon/INFE12-Jobs/Tuomi=AI%20on%20Education-EU-2018.pdf

53.

Venkatesh

Bala

(2008). Technology acceptance model 3 and a research agenda on interventions. Decision Sciences, 39(2), 273–315.

54.

Wang

Cheng

(2020). Cross-sectional studies: Strengths, weaknesses, and recommendations. Chest, 158(1), S65–S71.

55.

Wang

Y. M.

Wei

C. L.

Lin

H. H.

Wang

S. C.

Wang

Y. S.

(2024). What drives students’ AI learning behavior: A perspective of AI anxiety. Interactive Learning Environments, 32(6), 2584–2600.

56.

Wang

Y. Y.

Wang

Y. S.

(2022). Development and validation of an artificial intelligence anxiety scale: An initial application in predicting motivated learning behavior. Interactive Learning Environments, 30(4), 619–634.

57.

Yadav

(2025). Leveraging AI to enhance teaching and learning in education: The role of artificial intelligence in modernizing classroom practices. In Basha

J. S.

Alade

T. O.

Al Khazimi

M. O. A.

Vasudevan

Khan

J. B.

(Eds.), Optimizing research techniques and learning strategies with digital technologies (pp. 211–238). IGI Global Scientific Publishing.

58.

Zhou

Wang

(2018). The role of technology anxiety in the acceptance of health information technology. International Journal of Information Management, 38(1), 1–12. https://doi.org/10.1016/j.ijinfomgt.2018.02.012

Exploring Artificial Intelligence Anxiety Among Basic School Teachers in Ghana

Abstract

Plain Language Summary

Keywords

Introduction

Literature Review

Theoretical Foundation

Empirical Review

Methods and Materials

Sample and Study Design

Measure

Discriminant Validity

Confirmatory Factor Analysis for AI Anxiety Scale

Measurement Invariance

Data Collection Procedure

Data Analysis Strategy

Ethical Consideration

Results

Basic School Teachers’ Level of AI Anxiety

Differences in AI Anxiety Among Basic School Teachers Based on Gender

Discussion

Theoretical and Methodological Contributions

Practical Implications

Limitations of the Study and Suggestions for Future Research

Conclusions

Footnotes

ORCID iDs

Ethical Considerations

Funding

Declaration of Conflicting Interests

Data Availability Statement

References