In-service distance learners’ integration of generative artificial intelligence and emerging technologies: a comparative analysis with pre-service early childhood education teachers Open Access

Integrating digital technologies remains a central pedagogical challenge. Within this landscape, the emergence of generative artificial intelligence (GenAI) represents a disruptive force that can fundamentally alter teaching and learning processes (Li and Wong, 2023; Li et al., 2025; Varsik and Vosberg, 2024). For early childhood education (ECE), a field dedicated to establishing the foundations for lifelong development (Britto et al., 2017; Wong, 2022), the thoughtful incorporation of these advanced technologies is especially critical (Wong, 2026). As young children's lives are increasingly situated within digitally saturated environments, educators must adopt a responsive and informed pedagogical approach (Donohue, 2014; UNICEF, 2024).

Despite the urgency, the integration of technology within ECE often remains at a surface level. Cuban's (2001) observation that technology frequently serves as a “benign addition” (p. 67) continues to resonate, highlighting the challenge of moving beyond simple supplementation. To achieve meaningful integration, educators require a sophisticated form of digital competence, as defined by Fursykova et al. (2022) as the ability to understand, identify, and use digital technologies effectively in educational and professional activities. The potential of these tools thus remains unrealized when the educators who must wield them are not adequately prepared. Consequently, teacher education programs face a central challenge in bridging this gap.

While recent research has begun to investigate the attitudes of in-service early childhood educators toward these new tools (Wong et al., 2025), a critical gap remains in directly comparing educators' technology adoption patterns and perspectives across different professional development pathways. The issue is particularly salient for the growing number of practicing educators who engage in lifelong learning through open and distance education (ODE) programs (Perraton, 2010). These in-service educators are positioned to apply theoretical knowledge in real-time classroom settings. However, the specific demands of balancing full-time employment with academic study create a context for technology adoption that differs significantly from the traditional pre-service experience.

This study addresses this gap through a comparative analysis of two distinct cohorts of ECE students in Hong Kong pursuing the same qualification: distance-learning, in-service educators and full-time, pre-service teacher candidates. This design allows for a nuanced examination of the differences in technology use, perceived competencies, and perspectives between these two cohorts. The resulting insights highlight how ODE environments can uniquely connect theoretical technology integration with immediate classroom application. The study is guided by the following research questions:

The current technological wave in education is defined by a suite of emerging tools with transformative potential. GenAI encompasses systems that generate text, images, and media (Sengar et al., 2025). Key examples include large language models (LLMs) such as ChatGPT, Claude, Gemini, and DeepSeek (Wong, 2025). These models are trained on vast amounts of text data to generate human-like responses to answer questions or function as classroom tools (Meyer et al., 2023). Image generators such as DALL-E and Midjourney can create custom visual resources from simple text prompts. Educators can use these outputs for digital storytelling to personalize instruction and create culturally responsive content (Baskara, 2023). However, their deployment necessitates caution due to potential societal biases in training datasets (Navigli et al., 2023), risks to critical thinking (Gerlich, 2025), and data privacy concerns (Teo et al., 2024).

Beyond GenAI, other technologies are reshaping learning possibilities, particularly those that blend the physical and digital worlds. Immersive technologies such as augmented reality (AR) and virtual reality (VR) blend physical and digital worlds to support spatial reasoning (Al-Ansi and Fatmawati, 2023; Chan, 2025). Another key technology in this domain is the Internet of Things (IoT), which embeds connectivity and sensors into everyday objects to create hybrid learning opportunities. “Smart” toys and tools can then support playful exploration by linking physical actions to digital feedback (Ling et al., 2022). Collectively, research suggests these technologies can enhance learning by boosting student engagement and motivation (Mohamed et al., 2025). AR, for example, supports spatial reasoning by making difficult concepts more accessible (Olim and Nisi, 2020). However, significant barriers to widespread adoption remain. Common obstacles include high implementation costs, challenges of equitable and physical access for AR/VR headsets, and concerns about data security with IoT devices (Crogman et al., 2025; Ling et al., 2022).

The successful integration of innovative technologies hinges on teacher acceptance and competence (Chiu, 2022). This competence is not merely a technical skill; it is a multifaceted construct that requires educators to possess a deep pedagogical understanding of how to adapt technologies to specific educational settings, content areas, and available resources (Cabero and Barroso, 2016; Guzman et al., 2017). Furthermore, teachers' attitudes and perceptions are essential, as a positive outlook toward the value of technology is a critical factor influencing its successful adoption in the classroom (Lam and Ng, 2026; Wong et al., 2025).

However, many barriers can impede this integration, often leading to a cycle of low confidence and avoidance. Educators frequently feel overwhelmed, a feeling that often stems from systemic issues rather than individual shortcomings (Hernwall, 2016; Winter et al., 2021). For instance, a recent study of Hong Kong ECE educators found that over half lacked relevant training (58.8%), time (57.7%), and related electronic resources (52.6%) (Wong et al., 2025, p. 94). These individual teacher challenges can be understood as symptoms of a larger, systemic failure: a lack of “institutional readiness.” This concept posits that successful technology adoption hinges not on singular training events, but on a supportive institutional framework that provides the ongoing support necessary to overcome predictable barriers (Wong and Li, 2025, p. 2). Therefore, effective teacher preparation must equip educators not only with technical skills but also with the agency to advocate for and build these supportive environments.

The educational journeys of in-service educators in distance-learning programs and pre-service students in full-time programs are clearly different, yet they share a common starting point, making their comparison especially meaningful. Both groups in this study are enrolled in Bachelor of Education in Early Childhood Education (BEdECE) programs. All participants already possess registered kindergarten teacher status in Hong Kong, having completed a 2-year higher diploma in ECE with at least eight weeks of practicum. This shared foundation ensures the comparison is not between novices and experts, but between two different trajectories of further professional development.

A key component of the BEdECE program for both cohorts is the fulfillment of an additional four weeks of full-time practicum, as mandated by the Hong Kong Education Bureau for degree completion. This shared requirement ensures that all students, regardless of their employment status, have a contemporary practical setting to apply their learning. However, their experience of this practice differs significantly.

The distance-learning students represent a classic model of in-service professional development through a flexible, open learning modality. After completing their diploma, these individuals enter the ECE workforce and pursue their BEdECE degree concurrently while being actively employed. For these practicing educators, the additional practicum requirement is typically fulfilled within their existing employment. They bring a wealth of daily classroom experience to their studies and are in a unique position to immediately apply and test new pedagogical strategies and tools (including emerging technologies) in their professional settings. This continuous feedback loop between theory and practice can cultivate a highly contextualized and nuanced approach to technology integration. At the same time, these educators operate under different constraints common in distance education. Literature consistently identifies time poverty and the challenge of balancing work-study commitments as defining features of this demographic (Çalışkan et al., 2024; Li, 2023; Li and Wong, 2025).

In contrast, the full-time students embark on a continuous academic trajectory, directly entering Year 3 of the full-time BEdECE program. These students often belong to “Generation Z,” a demographic cohort that has grown up with the Internet and digital media as integral parts of their lives and are often considered digitally native (Chang and Chang, 2023; Kohnová et al., 2021). Their university-based learning is punctuated by required practicum blocks, which place them in kindergarten classrooms for extended periods.

These distinct educational and professional contexts thus form the basis for a valid comparison of how technology integration is perceived and enacted. The structural dichotomy mirrors findings in other jurisdictions, such as the United States and the United Kingdom, where research has similarly highlighted the divergent nature of employment-based routes versus traditional university-based programs (Campbell-Barr et al., 2020; Schachner et al., 2025).

To guide this comparative study, a conceptual framework was developed to illustrate the interplay between educator characteristics, contextual factors, and their perceptions and use of emerging technologies in ECE. This framework synthesizes core constructs from the Technology Acceptance Model (TAM) (Davis, 1989) and integrates key variables relevant to the educational context, such as professional training and support. TAM posits that an individual's intention to use a technology is primarily determined by two beliefs: Perceived Usefulness (PU) and Perceived Ease of Use (PEOU). Our framework adapts and extends this model to suit the specific nuances of in-service and pre-service teacher populations (Figure 1).

The framework posits that an educator's profile, defined by their training modality (distance-learning/in-service vs. full-time/pre-service), age, and experience, serves as the primary predictor. The training modality is central, as it distinguishes between a flexible, work-integrated pathway characteristic of ODE and a traditional, immersive academic program. This profile is hypothesized to be associated with a set of mediating perceptual and contextual factors. These include the core TAM constructs of PU and PEOU, as well as Self-Perceived Digital Competence and Facilitating Conditions (e.g. training, institutional support, resources). Finally, these perceptual and contextual factors are proposed as the primary determinants of the behavioral outcome: Actual Use of Technology, measured by frequency and type. While this initial framework guides our quantitative analysis, the mixed-methods design also allows us to explore other emergent themes from qualitative data that may enrich or challenge this model. Unlike other TAM adaptations that formally add constructs such as “Perceived Concerns” upfront (e.g. Li et al., 2025), our approach uses the qualitative component to allow these factors to emerge organically from the professional context.

A total of 288 students enrolled in BEdECE programs at a university in Hong Kong were recruited. As established in Section 2.3, participants were drawn from two distinct cohorts representing different professional development pathways: 118 distance-learning (in-service) students and 170 full-time (pre-service) students.

As detailed in Table 1, the cohorts differed significantly: distance learners were older, experienced professionals (mean age = 33.2), whereas full-time students were younger, pre-service candidates (mean age = 21.5). Both were predominantly female.

A self-developed questionnaire was designed to operationalize the constructs of the conceptual framework and address the study's research questions (Table 2). The instrument was created in Chinese to ensure clarity and accessibility for all participants in the Hong Kong context. To establish content validity, the initial draft of the questionnaire was reviewed by two senior academics with expertise in ECE and educational technology. Following their feedback, the instrument was pilot-tested with a small group of 4 ECE students (2 pre-service, 2 in-service) who were not a part of the final study sample. This pilot test served to refine the wording of the questions for clarity, check the logical flow, and confirm the estimated completion time. Furthermore, the internal consistency of the 13-item perceived benefits scale was evaluated, yielding a high level of reliability (Cronbach's α = 0.92).

The questionnaire was administered online using Qualtrics (https://www.qualtrics.com). A key feature was its responsive design, which presented subsequent questions based on a participant's prior answers. The logic ensured that individuals were asked only about the technologies they had actually used, thereby enhancing data relevance and reducing participant burden.

Ethical approval was granted by the university's Research Ethics Committee. Participants were recruited through announcements from placement coordinators and supervisors, as well as in-class invitations. The online questionnaire included an information sheet detailing the study's purpose and confidentiality measures. Electronic informed consent was obtained prior to participation, which was voluntary and took approximately 10 min.

A mixed-methods approach was used. Quantitative data were analyzed using IBM SPSS Statistics (Version 29). Descriptive statistics (frequencies, percentages, means, and standard deviations) were calculated to summarize the cohorts' responses. For comparative analysis, independent samples t-tests were used to compare mean scores on continuous variables, while Chi-square tests were used to examine associations between categorical variables. Pearson correlation analyses were performed to explore the relationships among continuous variables. A p-value of < 0.05 was set as the threshold for statistical significance. Textual responses to the open-ended question from 49 participants were subjected to thematic analysis (Braun and Clarke, 2006). The process involved familiarization, systematic coding by a trained research assistant, with the lead researcher validating the codes and themes for consistency. The codes were then collated into overarching themes, focusing on comparing perspectives between the distance-learning and full-time cohorts.

The results are presented in sequence to answer the research questions, providing a clear, logical progression of the findings.

Approximately half of the distance-learning students (50.85%, n = 60) and full-time students (44.71%, n = 76) reported never using GenAI or other emerging technologies in their teaching experience. Among those who reported use, adoption was often limited to a single tool; only 44.83% (n = 26) of experienced distance learners and 42.55% (n = 40) of experienced full-time students used more than one type of technology.

GenAI chatbots were the most commonly used technology (26.27% of distance learners; 35.29% of full-time students) (Figure 2). Among these adopters, approximately 80% in both cohorts reported monthly use. However, daily usage intensity differed: 13% of distance-learning adopters used chatbots daily, over four times higher than the 3% of full-time counterparts.

The primary purpose of using GenAI chatbots was designing lesson plans (54.84% for distance-learning; 61.67% for full-time). Their secondary priorities, however, diverged. For distance-learning students, the next most common purpose was setting learning goals (41.94%), followed by creating teaching materials (32.26%). A smaller but notable portion of users in both groups also used the tools to manage teaching materials and learning resources (16.13% for distance-learning; 18.33% for full-time).

Beyond GenAI chatbots, the IoT and GenAI image generators were the next most utilized tools, with inverted rankings between the groups. For distance-learning students, the second-most-utilized tool was IoT (17.80% adoption), with over half of these users reporting use at least once a month. Their third-most-prevalent technology was GenAI image generators (12.71%), with over 85% of users engaging with the tool at least once a month. In contrast, the pattern was reversed for full-time students, for whom GenAI image generators were their second-most adopted technology (14.71%), with almost 90% of these users reporting monthly or more frequent use. Their third-most-common tool was IoT (9.41%), with 75% of users reporting using it at least monthly.

A clear distinction in purpose was observed for these technologies. The primary purpose of using GenAI image generators for both cohorts was to create teaching materials (66.67% for distance-learning; 72% for full-time). On the other hand, the main purpose of using IoT was to increase student engagement (57.41% for distance-learning; 50% for full-time) and to provide multi-sensory learning opportunities (28.57% for distance-learning; 50% for full-time).

Immersive technologies were the least adopted by a significant margin. For distance-learning students, only 3.39% had used VR, and 2.54% had used AR. The adoption rates were even lower for full-time students, at 1.76% for VR and 1.18% for AR.

A greater proportion of distance-learning students (22.9%) reported receiving specific training in GenAI and other emerging technologies than full-time students (10.0%). This situation contrasts with attendance in a general information technology (IT) course. For full-time students, this course was a required part of the curriculum, with 68.8% having attended. For distance-learning students, for whom the course was optional, only 21.8% attended.

In self-assessments of technology usage (Table 3), independent samples t-tests revealed that distance-learning students scored significantly lower than full-time students in perceived ease of use (Ms = 4.58 vs. 5.11, p = 0.047) and the necessity of using technology in teaching (Ms = 5.66 vs. 6.19, p = 0.032). Distance-learning students also reported lower mean proficiency scores (Ms = 4.14 vs. 4.47), but this difference was not statistically significant (p = 0.217).

To investigate the impact of formal training, further analysis revealed a sharp contrast between the cohorts. For the distance-learning students, the choice to attend the optional IT course was associated with significantly higher self-perceived proficiency (p < 0.001) and greater ease of use (p = 0.02), as shown in Table 4. Importantly, this boost in confidence appears to be linked to actual behavior. A Chi-square test found a statistically significant, moderate association between attending this optional course and applying GenAI and emerging technologies in teaching (χ²(1, n = 101) = 7.76, p = 0.005, Cramer's V = 0.277). In contrast, this effect was absent in the full-time cohort. For these students, the mandatory IT course had no statistically significant impact on their perceived proficiency, ease of use, or necessity of use (all ps > 0.05), nor was it associated with the practical application of GenAI and emerging technologies in their teaching (χ²(1, n = 170) = 0.01, p = 0.92).

Furthermore, a weak but statistically significant negative correlation was observed between age and agreement on the necessity of using technology in teaching (Pearson's r = −0.21, p = 0.02) within the distance-learning cohort. The correlation indicates that, within this group of in-service educators, younger participants were more likely to agree that technology is a necessity for teaching compared to their older colleagues.

Regarding benefits (Table 5), a strong consensus emerged on several fronts. Over 90% of both distance-learning and full-time students agreed that enhancing student classroom engagement is a significant benefit. For distance-learning students, the next-highest perceived benefits were that technology helps create teaching materials (89.74%) and provides multi-sensory learning experiences (89.66%). Nearly 96% of full-time students also agreed that these technologies offer multi-sensory learning opportunities.

However, alongside these benefits, participants identified significant challenges that hinder effective integration (Table 6). The cohorts reported different hierarchies of challenges. Distance-learning students primarily struggled with having “no time to learn about technology” (53.39%) and a “lack of related electronic resources and equipment” (45.76%). In contrast, the biggest challenge for full-time students was the “complexity and difficulty of applying technology” (50.59%), followed by a “lack of relevant training opportunities” (40.59%).

To add context to the quantitative findings, analysis of the written responses revealed three key themes that illuminate the perceptions of benefits and challenges (RQ3): (1) the conviction in the primacy of the human educator, (2) the systemic and practical challenges hindering adoption, and (3) concrete recommendations for moving forward.

A predominant theme was the conviction that GenAI and emerging technologies cannot supplant the essential human role in ECE. Respondents argued that the technologies' lack of emotion and nuanced understanding makes them incapable of imparting moral values, positioning these tools as functional assistants rather than educator replacements.

One distance-learning student noted AI's “fixed and rigid” procedure cannot adapt to children it does not “personally know” (Student D38). Another respondent elaborated on this distinction, suggesting that while GenAI might extract “the theoretical principles from a story,” the emotional and moral dimensions “still require a traditional teacher to convey them properly” (Student D66).

The primary benefit of GenAI was therefore seen in the preparatory phase. One participant noted the value of using ChatGPT to get “ideas for new lesson plans or activity themes” (Student F21), highlighting its role as a brainstorming partner. Some also saw benefits in direct, controlled application. For instance, one distance-learning student described a positive experience using GenAI-driven projections for interactive physical games, noting that the “children really enjoy this and find it very engaging” (Student D03). The example illustrates a perceived benefit where GenAI can create novel, multi-sensory activities under the teacher's guidance.

Respondents detailed a range of practical challenges that explain the constrained and hesitant usage of new technologies. These went beyond pedagogical philosophy to include significant systemic barriers.

A primary issue was a simple lack of resources and infrastructure, with one distance-learning student stating, “The school does not have AI tools yet” (Student D09). There were also doubts about the developmental appropriateness of existing technology. The same respondent shared a personal experience with VR, noting, “I feel dizzy after using it, so I feel it is not suitable for use in early childhood education” (Student D77). Another participant raised the pedagogical concern that too many “online activities” could lead to children getting “bored from constantly facing a computer or an iPad” (Student F11), diminishing engagement rather than enhancing it.

This combination of factors contributes to a reliance on self-sufficiency. Several respondents expressed a preference for their own professional judgment, turning to GenAI and emerging technologies only as a last resort. One distance-learning student explained:

In direct response to these challenges, respondents offered several key recommendations to help realize the potential benefits of GenAI and emerging technologies, focusing on professional development and systemic support.

A critical need identified was for targeted teacher training that moves beyond basic operation to pedagogical application. As one participant wrote, “Although we know how to operate each GenAI tool, it would be beneficial if the school could offer more workshops on which specific tool we can use in our teaching” (Student F84). This was seen as a particular challenge for currently practicing educators, as one respondent explained that “in-service teachers may have less exposure to this area” compared to students not yet in the workforce (Student D38). One distance-learning student stated plainly:

In addition to training, participants called for government intervention. Some suggested financial subsidies similar to existing textbook assistance schemes to help schools purchase technology. A distance-learning student advocated for earmarked funding, asking, “Could the government provide more subsidies … a sum of money that they are required to spend specifically on AI, such as purchasing AI equipment?” (Student D09).

This study sought to compare the integration of GenAI and emerging technologies between two distinct cohorts of ECE teacher candidates: in-service educators studying via distance learning and pre-service candidates in a full-time program. The findings reveal a complex picture in which the specific professional and educational contexts of these groups are associated with different patterns of technology adoption, self-perception, and perceived barriers. The discussion interprets these findings in relation to the study's research questions and conceptual framework, with a particular focus on the implications for ODE.

Despite distinct usage patterns, a substantial proportion of participants across both cohorts have not yet adopted these tools. Such hesitation across both pre-service and in-service pathways underscores a broader reality: despite the rapid proliferation of educational technologies, their actual integration into everyday ECE practice remains limited.

For educators who do utilize these tools, the divergence in their application suggests two distinct modes of engagement. The broader adoption of GenAI chatbots by full-time students is consistent with a pattern of digital exploration, characteristic of a generation comfortable experimenting with new tools (Chang and Chang, 2023). In contrast, the deeper, more frequent daily integration by a smaller group of distance-learning students points to a highly pragmatic approach. Such focused application resonates with the challenge of moving technology beyond a “benign addition” (Cuban, 2001, p. 67). For these in-service educators, GenAI primarily serves as a targeted solution to immediate professional needs, such as lesson planning and setting learning goals. Rather than a mere behavioral quirk, the intensity of use among distance learners indicates that technology is being pulled into practice by necessity — a dynamic central to the value proposition of continuing professional education through flexible learning (Perraton, 2010).

A paradox emerges regarding digital competence and professional development. While extensive in-service experience might seem to correlate with settled practices, the distance-learning cohort demonstrated a more proactive approach to seeking targeted training. Confronting real-world classroom complexities daily appears to foster a more critical, realistic self-assessment among these educators. Rather than indicating a lack of ability, their lower self-perceived proficiency likely reflects a deeper understanding of the practical hurdles involved in technology integration. Recognizing these challenges prompts them to pursue tailored training, aligning perfectly with Fursykova et al.’s (2022) definition of digital competence as the capacity to utilize technology for effective professional activity, rather than mere technical skill.

The power of self-directed learning becomes especially evident when examining the differing associations between IT course attendance and technology use across the two groups. For distance learners, self-selecting into an optional IT course was associated with higher confidence and actual classroom application. Conversely, the mandatory, curriculum-embedded course completed by the pre-service cohort showed no such behavioral association. Such a sharp contrast reinforces the foundational principles of andragogy: adult learners are most receptive to problem-centered instruction that is immediately applicable to their professional lives (Knowles, 1990). The data provide compelling evidence that effective professional development for in-service teachers relies less on curriculum mandates and more on providing access to need-driven, self-initiated learning opportunities.

Furthermore, the observed negative correlation between age and the perceived necessity of technology among in-service teachers highlights the influence of established pedagogical beliefs on adoption patterns (Chiu, 2022). Interpreting these age-related differences, however, requires caution due to the significant heterogeneity within the distance-learning population. The cohort comprises educators with vastly different career lengths; veterans with decades of successful, established teaching methods may naturally feel less need for disruptive tools compared to newer teachers still refining their practice. Acknowledging such diversity is important for ODE programs, as they must design support structures that cater to a wide spectrum of pre-existing professional identities and attitudes.

While both cohorts agreed on technology's benefits for student engagement, their identified challenges clearly reflected their different circumstances. For the distance-learning cohort, a lack of time emerged as the primary barrier, representing a classic constraint for in-service professionals undertaking flexible study (Çalışkan et al., 2024; Li, 2023; Tang et al., 2025). Coupled with reported deficits in resource and school support, these constraints offer an empirical illustration of the “institutional readiness” concept (Wong and Li, 2025). The evidence suggests that the primary obstacles for these practicing educators are systemic rather than personal, stemming from a lack of an enabling environment in their workplaces. Conversely, full-time students primarily struggled with the complexity of the technology itself, pointing to an inherent technical and pedagogical learning curve. Recognizing these differing priorities is essential for designing effective teacher education: pre-service training should focus on demystifying technology, whereas ODE programs must additionally address time management and provide guidance on advocating for resources.

The qualitative findings add crucial depth to these barriers, particularly through the emergent theme of the “primacy of the human educator.” Such a conviction serves as an important boundary condition for the TAM in the ECE context. Educators in both groups perceive usefulness (PU) and ease of use (PEOU) not as absolute measures, but as constructs subordinate to their professional ethics and pedagogical philosophy. This view helps explain the hesitant and self-reliant attitudes expressed by participants, who often view GenAI as a tool of last resort rather than a primary resource.

This emphasis on human primacy offers a valuable counterpoint to recent research on undergraduates in general. For instance, Li et al. (2025) found that “Perceived Concerns” about AI did not significantly predict usage frequency and concluded that immediate perceived benefits outweighed ethical or practical worries for that demographic. Our study suggests a different reality for professional educators. The nature of the ECE profession appears to imbue these concerns with pedagogical and ethical weight, making foundational beliefs a far more salient factor in adoption decisions than they might be for the general student population. Consequently, the professional context emerges as a critical moderator of technology acceptance, a factor that generic adoption models often overlook.

The current research offers important theoretical and practical implications, particularly for the design and delivery of ODE. From a theoretical standpoint, adapting the TAM proved useful, yet the framework requires enrichment to capture the nuances of teacher education. Specifically, the “Educator Profile,” which highlights the dichotomy between in-service and pre-service pathways, acts as an important moderator. Qualitative insights further indicate that integrating constructs such as “Pedagogical Beliefs” or “Perceived Professional Role” would provide a more comprehensive model of technology acceptance among educators. The necessity for such context-specific variables becomes especially apparent given the limited explanatory power of generalized TAM-based models. For example, Li et al. (2025) reported that their model, which included benefits, ease of use, knowledge, and concerns, explained only 23% of the variance in AI usage frequency among general undergraduates. Such low variance strongly implies that unmeasured factors, namely the professional and pedagogical convictions identified here, play a decisive role in adoption behaviors.

On a practical level, the comparative analysis provides actionable guidance for universities worldwide, especially those delivering distance education to working professionals beyond the local context. First, ODE providers should actively mitigate the acute time poverty experienced by their students. Designing highly flexible, modular, and just-in-time training opportunities aligns well with the principles of Agile-Blended Learning (see: Li, 2023; Li and Wong, 2025; Wong, 2024; Wong and Li, 2025; Tang et al., 2025). Second, rather than solely focusing on technical operation, the curriculum should equip in-service teachers with strategies to integrate these tools within the realistic constraints of their classrooms, including advocacy for institutional support. Third, the demonstrated value of targeted, self-selected training suggests that ODE programs should pivot toward curating specialized workshops focused on the direct pedagogical application of specific tools such as GenAI, rather than relying solely on broad, mandatory IT courses.

Despite these contributions, several limitations warrant consideration. The cross-sectional design precludes the establishment of causal relationships; therefore, the observed differences represent associations rather than direct outcomes of the respective learning modality. Also, reliance on self-reported data introduces the potential for social desirability bias. The specific cultural and structural context of Hong Kong's ECE sector may also limit the generalizability of the findings to other educational systems. Finally, the unequal cohorts necessitated specific statistical adjustments, a factor that requires careful consideration when interpreting the comparative outcomes.

The discourse on educational technology is often dominated by the capabilities of the tools themselves. This study, however, shifts the focus back to the educator. The findings reveal that the path to meaningful technology integration is not merely about abstract training but about the context of daily professional practice. Such evidence challenges the linear assumption that teachers must first be taught about technology before they can use it effectively. Instead, the data suggest a more dynamic reality, particularly for in-service educators: the classroom is not a passive site for applying pre-learned skills but an active crucible that forges the purpose and utility of new tools.

These insights offer a significant opportunity for ODE. Instead of viewing physical separation from a traditional campus as a limitation, ODE programs should embrace their unique position at the intersection of theory and practice. The evidence indicates that the most effective support ODE can provide may not lie in mandatory courses that imitate a traditional curriculum. The greatest value instead resides in creating a rich ecosystem of flexible, just-in-time, and need-based learning opportunities that enable self-directed professionals to access knowledge exactly when they need it. The challenge for ODE, therefore, is not merely to teach technology but to innovate its pedagogical models to ensure technology enhances professional practice rather than dominating it. The goal is not to build technologically saturated classrooms, but to cultivate pedagogically wise teachers.

This study was reviewed and approved by the Research Ethics Committee of Hong Kong Metropolitan University (Reference No. HE-SOL-SPG2023/01).

The work described in this paper was fully supported by the Hong Kong Metropolitan University Small Project Grant (No. SOL-SPG2023/01). The authors wish to express their sincere gratitude to Miss Jasmine Suet Man Chau for her invaluable research assistance throughout this project. We are also deeply grateful to all the students who participated in this study. Their willingness to share their time and perspectives was essential to the completion of this work.