Impact of AI technology adoption on workload management:

A pilot study with special references to private university lecturers in Colombo district

K.S. Nethmini

Undergraduate, Faculty of Management, Horizon Campus, Sri Lanka.

 K.T.D. Jayathilake

Senior Lecturer, Faculty of Management, Horizon Campus, Sri Lanka.

Correspondence: nethminiwijesinghe09@gmail.com

Abstract

Artificial intelligence is becoming increasingly common in higher education, changing how academic staff handle teaching and administrative work. However, very few studies have examined its effect on instructors at private universities in the Colombo District, particularly the possibility that AI could add to their workload through extra learning and verification demands. Therefore, this pilot study examines the impact of AI adoption, specifically AI-enabled task automation, AI-assisted content creation, and digital communication tool integration, on lecturer workload management, weighing its benefits and drawbacks. Following a positivist philosophy and a deductive approach, data were collected from 15 private university lecturers in the Colombo district using a 30-item, 5-point Likert scale questionnaire distributed through Google Forms, selected via simple random sampling, and analysed in SPSS through reliability testing, correlation, and multiple regression. All three independent variables correlated significantly and positively with workload management (r = .626–.829, p < .05) and together explained 75.8% of the variance (adjusted R² = .693, F(3,11) = 11.515, p = .001). Content creation emerged as the strongest and only statistically significant unique predictor (β = .572, p = .023), while task automation (p = .649) and digital communication (p = .136) were weaker, likely because of the overlap between the constructs rather than multicollinearity (VIF = 1.73–2.13). This indicates that H2 was fully supported, whereas H1 and H3 were only partially supported. Given the small pilot sample, single-campus context, age concentration, and some departure from residual normality, these findings are preliminary and specific to the lecturers in Colombo district. However, they still indicate that AI adoption, particularly through content creation tools, meaningfully supports workload management, offering an early contribution to how technology adoption in Sri Lankan higher education is understood, and suggesting that private universities should prioritise AI training in content creation and communication tools while larger, multi-campus research is conducted to confirm how well these results generalise.

Keywords: AI Adoption, Content Creation, Digital Communication, Task Automation, Workload Management.

JEL: O33, J24

1. Introduction

Artificial intelligence is no longer a novelty in higher education; it has quietly become part of how lecturers do their jobs. Teaching staff now use AI to prepare lesson materials, mark assignments, design slides and quizzes, and keep in touch with students through chat tools and learning platforms. This isn't happening only in large, well-funded universities. In Sri Lanka, private universities and other degree-awarding institutes recognised by the University Grants Commission have grown rapidly over the past several years, and many of them run a wide range of programme with a fairly lean academic staff (University Grants Commission Sri Lanka, 2023). One knock-on effect of this is that lecturers end up carrying a mix of duties, teaching, research, administrative paperwork, and constant student support, which can pile up quickly. How well this load is managed matters, both for the quality of a lecturer's work and for their own well-being, which is the starting point for this study.

AI-enabled task automation, AI-assisted content creation, and digital communication tools are usually discussed as a solution to this kind of pressure. Software can now grade routine work, generate quiz questions, draft slides, and handle a good chunk of day-to-day correspondence with students, and in theory, that should free up hours a lecturer can put toward better teaching or their own research. Ejaz et al. (2025) found exactly this kind of pattern among lecturers in Pakistan, where AI adoption was driven mainly by how much it improved performance and how easy it was to pick up. However, Liu et al. (2026), looking at teachers in China, added a more cautious note: continued use only held up when the AI tool actually matched what a teacher needed for real classroom tasks, not just when it was available. So the promise of AI easing workload seems to be real, but conditional, and it isn't yet clear how that condition plays out specifically for lecturers working in Sri Lankan private universities, where training, infrastructure, and institutional support can look quite different from what's available in state universities or in universities overseas.

There is a fair amount of tension in the literature regarding whether AI actually helps with the workload. Some studies report that it reduces administrative pressure and protects lecturers from burnout by absorbing repetitive tasks (Hashem et al., 2024). Others paint a less flattering picture of the future. Monahan and Burlacu (2024), for instance, found that 77% of educators using AI tools actually reported a rise in their workload, not a fall, largely because of the stress of learning unfamiliar systems without adequate training. This is a genuinely awkward finding because it suggests that simply pushing more AI tools into a university without thinking through implementation could easily backfire rather than help. Gardan et al. (2025) offer part of an explanation here: they found that familiarity with AI shapes how useful people perceive it to be, which then shapes attitude and how likely someone is to keep using it. In other words, the same tool can look like a burden to one lecturer and a genuine time-saver to another, depending largely on how comfortable they already are with it, a pattern also picked up by Atherton et al. (2026), who found that a lecturer's personal view of AI predicted actual use more strongly than any institutional push toward adoption. Barham (2026) adds a further caution worth keeping in mind from the outset: even where AI saves time, over-reliance on it can quietly erode a lecturer's own critical thinking if it is not paired with proper training and some ethical guardrails.

Stepping back, there is a gap in where all this research has actually been done. Most of it comes out of state university systems or out of educational contexts in wealthier, Western countries. Very little has looked specifically at private university lecturers in developing countries, and Sri Lanka fits squarely into that blind spot. On top of that, a lot of the existing work leans heavily on quantitative surveys that are good at spotting broad patterns but don't say much about what it actually feels like, day to day, for a lecturer trying to fold a new AI tool into an already full schedule. This study tries to respond to that gap, at least in a small way, by looking specifically at how AI adoption relates to workload management among lecturers at private universities in the Colombo district.

It is worth noting that this is a pilot study and not the final word on the subject. The point here is not to produce sweeping, generalisable conclusions immediately, but to try the research approach on a smaller group of lecturers, see how the variables behave, and catch anything that needs fixing before running the full-scale study. That's a fairly standard, sensible way to start exploratory work in an area where local evidence is still thin, and the findings later in this paper should be read with that in mind.

With that context set, this study zeroes in on three specific areas of AI use that keep coming up in the literature: AI-enabled task automation, AI-assisted content creation, and digital communication tool integration. The aim, put simply, is to examine how adopting these technologies relates to workload management among lecturers at UGC recognized private universities in Sri Lanka. This breaks down into three more specific goals: to look at how task automation affects lecturer workload, how content creation tools affect it, and how digital communication tool integration affects it. Splitting the question up this way lets the study move past a vague "does AI help or not" and get into where, specifically, its effects show up most and least, which matters given how differently these three areas performed in studies like Guo and Wei (2025), who describe AI as capable of acting as either a demand or a resource on a lecturer's time, depending on how it's actually used.

The significance of this work is fairly practical. For university management, the findings can point toward where AI genuinely reduces staff pressure and where more caution or training is needed before rolling something out further. Lecturers today are already balancing teaching, marking, research, admin, and near-constant student communication, and when that load gets too heavy, it shows up in both performance and personal well-being. Zooming in on task automation, content creation, and communication tool integration gives a clearer picture of which of these actually earn their keep for lecturers in this setting, rather than assuming AI helps across the board. The results could also be useful to the University Grants Commission and other policymakers working on guidelines for AI use in higher education, since decisions about training budgets, software investment, and institutional policy tend to work better when they are grounded in local evidence rather than borrowed assumptions from elsewhere. And because most of the existing research on this topic comes from outside Sri Lanka, even a modest pilot like this one adds something to the record, and it can act as a starting point for other researchers wanting to look further into AI adoption, academic workload, or lecturer well-being, whether in Sri Lanka or in similar developing-country settings.

2. Literature Review

There is a growing body of research on how AI is used in academic work, and a few clear patterns emerge once you start pulling the studies together. Broadly, people seem to adopt AI tools when the tools genuinely make their work easier and do not require much effort to learn, and the tools tend to help most when they are well matched to what the person needs to get done day to day. This basic idea runs through all three areas this study looks at.

Starting with AI-enabled task automation, meaning AI used for things like grading, attendance records, and timetabling, Ejaz et al. (2025) looked at lecturers in Pakistan and found that they picked up AI mainly because it made them more effective and wasn't hard to use. Liu et al. (2026) added something worth paying attention to here: teachers in China only kept using an AI assistant over the long run if it actually matched their real teaching tasks. If it didn't fit well, they simply stopped using it, even when the tool itself was perfectly capable. In summary, these two studies suggest that automation can genuinely reduce routine workload only when the tool fits the job it's meant to help with, not just because it exists.

Next is AI-assisted content creation, which includes preparing lecture notes, slides, quiz questions, and academic writing with AI's help. Gardan et al. (2025) found that lecturers who were more familiar with these tools saw them as more useful, which led to a better attitude toward them and a stronger intention to keep using them. Atherton et al. (2026) took this a step further, finding that a lecturer's own personal opinion of AI predicted whether they'd actually use it far more strongly than any push from the university itself. Thus, individual attitude seems to matter more here than institutional policy. However, Barham (2026) raised a fair warning: the benefits of using AI to create content depend a lot on doing it ethically and with proper training, because leaning on it too heavily can quietly wear down a lecturer's own critical thinking over time. Worth keeping in mind, since it means these tools aren't automatically a good thing; how responsibly they're used matters just as much as whether they're used at all.

The third area is digital communication tool integration, meaning things like messaging apps, video calls, chatbots, and LMS notifications becoming part of everyday academic work. Demuyakor et al. (2026) found that these tools improved collaboration between staff and students, but they also came with a real cost, since constant messaging blurred the line between work hours and personal time. Kazmi and Irshad (2025) were more upbeat, finding that these tools boosted productivity by improving job satisfaction, especially when communication felt manageable rather than overwhelming. Cianci et al. (2024) helped explain why these findings did not quite agree with each other: they found that having support and some control over communication tools genuinely helped, but feeling pressured to reply immediately caused exhaustion. So digital communication tools seem to help only when lecturers have some say over when and how they respond, rather than feeling like they need to be available all the time.

Workload management itself is the last piece here, and it basically comes down to how well a lecturer's work, teaching, marking, research, admin, and student support is planned, shared out, and finished within the time they actually have. Pasha et al. (2025) found that digital fatigue and mental overload clearly hurt performance, but that good workload management can soften those effects quite a bit. Guo and Wei (2025) had perhaps the most useful finding for this study: AI can either add pressure to a job or genuinely help with it, depending entirely on how it's brought in and managed. Where workload isn't managed well, the outcome tends to be stress, burnout, and weaker teaching. Where it is managed well, AI seems to be absorbed more comfortably into a lecturer's routine instead of adding to the strain.

Pulling all this together, the pattern across the three areas is fairly consistent. Task automation, content creation, and communication tools all have real potential to ease a lecturer's workload, but none of them help unconditionally. Automation only helps when it actually fits the task at hand. Content creation only helps when it is used responsibly, and lecturers have had proper training. Communication tools only help when lecturers have some control over how and when they engage with them. In every case, the same tool that could save time can just as easily add to the pressure if it is introduced carelessly.

Given all this, and given how little direct research exists on Sri Lankan private university lecturers specifically, this study proposes three hypotheses to test on a small, exploratory scale. H1 is that AI-enabled task automation has a positive effect on workload management among private university lecturers. H2 is that AI-assisted content creation has a positive effect on workload management among private university lecturers. H3 is that digital communication tool integration has a positive effect on workload management among private university lecturers. Since this is a pilot study with a small sample from the Colombo district, these hypotheses are not meant to produce final, sweeping conclusions. The goal is to obtain early evidence that can point the way toward a larger, more representative study.

3. Hypothesis

H_0: AI – enabled task automation has no significant impact on workload management

H_1: AI – enabled task automation has a significant positive impact on workload management

H₀: AI – assisted content creation has no significant impact on workload management.

H₂: AI – assisted content creation has a significant positive impact on workload management.

H₀: Digital Communication tool integration has no significant impact on workload management.

H₃: Digital Communication tool integration has a significant positive impact on workload management.

4. Methodology

This study takes a positivist approach, working from the idea that reality can be observed and measured objectively rather than shaped by the researcher's own views. This fit was practical rather than purely philosophical: all four variables here, AI-enabled task automation, AI-assisted content creation, digital communication tool integration, and workload management, needed to be turned into structured Likert-scale questions and run through statistical tests, so a positivist stance made the most sense.

The approach is deductive. The three hypotheses were not pulled out of thin air; they came out of the literature reviewed earlier on workload theory and AI adoption and were set before any data were collected. The data collected from lecturers were then used to test those hypotheses, not to build new theory from scratch.

A survey was used as the strategy here. A structured questionnaire, sent out through Google Forms, provided a quick and inexpensive way to obtain standardised, numerical answers from lecturers, which was necessary for the later correlation and regression work.

Because this was a pilot study, the sample was small by design and should not be mistaken for a representative one. Fifteen lecturers from a private university in the Colombo district participated. They were chosen through simple random sampling; thus, each lecturer had an equal chance of being selected, the same logic that would apply in a full-scale version of this study, but on a much smaller scale. Fifteen people are obviously not enough to say anything definitive about lecturers in the Colombo district generally; however, using proper random selection even at this stage kept the process consistent with how the eventual larger study would be run. The unit of analysis was individual lecturers.

The questionnaire comprised 30 items, built around how lecturers perceive their academic and administrative tasks, and used a 5-point scale from strongly disagree to strongly agree. Alongside a few demographic questions, it covered each of the three independent variables and the one dependent variable, workload management. Google Forms made distribution simple and kept the responses in a format that could go straight into statistical analysis.

All analysis was done in SPSS. Descriptive statistics were used first to obtain a basic picture of the 15 respondents. Cronbach's alpha was used to check whether the questions meant to measure each variable were actually holding together consistently, which matters even more at this early stage, since catching a weak or inconsistent item now is the whole point of running a pilot before scaling up. Multiple regression was the main tool used to test the hypotheses, looking at how task automation, content creation, and digital communication tool integration each relate to workload management, both together and on their own.

The reason for setting things up this way goes back to a gap identified earlier in the literature review: there is simply not much local, evidence-based research on how AI adoption helps or hurts workload management for private university lecturers in Sri Lanka. Running this first as a small, single-campus pilot makes it possible to iron out any problems with the questionnaire or the analysis before committing to a bigger, multi-university sample down the line. Given how small and localised this sample is, the results should be read as an early, exploratory check on the research design, not as a final answer.

5. Findings/Results

Before conducting the hypothesis testing, the questionnaire was checked to ensure that it measured things consistently, using Cronbach's alpha. The scores for AI-enabled task automation, AI-assisted content creation, digital communication tool integration, and workload management all returned at acceptable levels; therefore, even with a small pilot group, the questionnaire held up reasonably well. Basic descriptive statistics were also run on the demographic data to get a feel for who responded, which is discussed later under the limitations section.

The main analysis showed that all three areas–task automation, content creation, and digital communication tools–were significantly and positively linked to workload management. The correlations ranged from r = .626 to r = .829, all significant at p < .05. These were checked again using Spearman's correlation, which handles small samples and non-normal data better, and it supported the same pattern. In plain terms, lecturers who said they used AI more in these three areas also tended to report managing their workload better, and that held true regardless of which test was used to check it.

Moving on to regression, all three variables were placed into one model to determine how much they explained workload management, both as a group and individually. Together, they accounted for 75.8% of the variation in workload management, with an adjusted R² of .693. The adjusted number is the one worth trusting more here, since it corrects for the small sample and the number of predictors, and .693 still points to a strong relationship. The model overall was statistically significant, F(3,11) = 11.515, p = .001, so this isn't just noise, the three variables together genuinely predict workload management outcomes for the lecturers who took part.

Looking at each variable on its own inside the regression tells a slightly different story than the correlations alone suggested. Content creation came out as the strongest, and the only one that stayed statistically significant once the other two were accounted for, with a beta of .572 and p = .023. So, even after controlling for automation and communication tools, content creation still made its own distinct difference to how well lecturers managed their workload, which fits with what Gardan et al. (2025) and Atherton et al. (2026) found, that familiarity and personal attitude toward AI strongly shape how much people actually get out of using it. Task automation, on the other hand, didn't hold up as a significant predictor on its own (p = .649), and neither did digital communication tool integration (p = .136).

It would be a mistake to read those last two results as meaning that automation and communication tools do not matter at all. The VIF values across the three predictors ranged from 1.73 to 2.13, well within an acceptable range; therefore, multicollinearity is not the problem, the variables are not just too similar to distinguish mathematically. What's more likely going on is that these things overlap in how lecturers actually experience them. Task automation and communication tools probably feel, to the people answering the survey, closely tied up with how they use content creation tools, so once content creation's effect is pulled out, there isn't much unique signal left for the other two to show, especially with such a small sample. This lines up with something Guo and Wei (2025) pointed out, that AI can end up feeling like either extra pressure or genuine help depending on how it's used, which fits the idea that lecturers may be experiencing all three of these tools as one connected way of working rather than as three separate things.

Coming back to the hypotheses, H2, that content creation has a positive impact on workload management, is fully supported by the data. H1 and H3, covering task automation and digital communication tools, are only partly supported. Both showed a strong positive correlation with workload management by themselves; however, neither held up as an independent predictor once all three were considered together. This is a useful thing to learn from a pilot study, not a disappointing one. This suggests that content creation deserves close attention in a larger follow-up study, while automation and communication tools might need to be looked at differently, maybe tested on their own or dug into more to understand how they overlap with content creation.

A few limitations matter here, and they are the reason this is called a pilot rather than a finished piece of research. The biggest one is the sample size, just 15 lecturers, which is well below what you'd normally want for a stable regression analysis; therefore, the exact numbers should not be taken too literally. The sample also came mostly from one campus and skewed toward a similar age group; therefore, what shows up here might just reflect that particular group rather than private university lecturers across Colombo district generally. In addition, checking the regression residuals showed that they were not perfectly normal, which often happens with small samples and means that the model, while still informative, is not fully stable yet. In summary, these limitations mean that the findings here are early and specific to this small group of Colombo district lecturers. They are best treated as a starting point for sharpening the questionnaire and research design before running a bigger, more representative study, not as a final answer about AI adoption and workload management across Sri Lankan private universities.

6. Conclusion and implications 

This pilot study examined the impact of AI technology adoption on workload management among lecturers working in private universities in the Colombo district, focusing specifically on three areas of AI use: task automation, content creation, and digital communication tool integration. Taken as a whole, the results show that AI adoption has a positive relationship with how well lecturers manage their workload, but this effect is not evenly spread across all three areas. Content creation emerged as the main driver of this positive relationship, standing out as the only variable that remained a statistically significant predictor once all three were considered together. Task automation and digital communication tool integration both showed encouraging positive correlations with workload management on their own, but their individual contributions became harder to separate out once content creation was accounted for in the same model. In short, AI appears to help lecturers manage their workload, and the way it helps most clearly, at least in this pilot, is through tools that support the preparation of lecture notes, slides, quizzes, and academic writing.

From a theoretical perspective, these findings offer a useful, albeit modest, contribution to understanding technology adoption in the context of Sri Lankan higher education. The strong role played by content creation fits comfortably within the Technology Acceptance Model discussed earlier in this paper, particularly the idea that a lecturer's own sense of a tool's usefulness and ease of use strongly shapes how much benefit they draw from it, as discussed earlier. Content creation tools are used frequently, directly touch a task that lecturers already spend a great deal of time on, and produce a visible, immediate reduction in effort, which likely explains why their effect was so clear even in a small sample. The weaker, less distinct results for automation and communication tools, meanwhile, connect back to the JD-R model referenced in the literature review, which treats the same technology as capable of acting as either a demand or a resource depending on how it is experienced. It is plausible that, in the minds of the lecturers surveyed, automation and communication tools are not experienced as separate, standalone influences on workload but rather as part of a broader, interconnected AI-supported way of working that overlaps with content creation. This suggests that future theoretical work on AI adoption in academic settings may benefit from treating these variables as more interrelated constructs rather than fully independent ones, an idea that this pilot study was well placed to surface precisely because of its close variable-by-variable analysis.

On the practical side, the findings offer fairly direct guidance for private university management in Sri Lanka, even allowing for the pilot nature of the study. Since content creation tools showed the clearest and strongest link to improved workload management, universities have a reasonably solid basis for prioritising training and support in this specific area first, rather than spreading limited training budgets thinly and evenly across every possible AI application. Lecturers who are more familiar and comfortable with AI-assisted content creation tools appear better positioned to manage their overall workload, which points toward practical steps such as structured workshops on using AI for lecture preparation, slide design, and assessment writing, along with clear institutional guidance on using these tools responsibly, in keeping with the ethical cautions raised by Barham (2026) about over-reliance weakening critical thinking. At the same time, the encouraging correlations found for task automation and digital communication tools suggest that these areas should not be neglected. Rather than assuming they matter less, universities should consider introducing them in a more deliberate, well-supported way, since earlier literature reviewed in this paper, particularly Cianci et al. (2024) and Liu et al. (2026), indicated that these tools only genuinely help when they are well matched to real tasks and when lecturers retain a reasonable degree of control over how and when they are used.

It is important to be honest about what this pilot study can and cannot claim. With a sample of only 15 lecturers drawn largely from a single campus and clustered around a similar age range, these findings cannot be treated as representative of private university lecturers across Colombo district, let alone Sri Lanka as a whole. The imperfect residual normality noted in the results section is a further reminder that the statistical model, while informative, is not yet fully stable. These limitations do not undermine the value of the pilot; they are, in fact, exactly what a pilot study is meant to reveal, giving early direction before committing resources to a larger data collection effort. What this pilot does convincingly show is that the overall research design works, the questionnaire produces reliable and interpretable data, and the three hypotheses are worth carrying forward into a bigger study.

Looking ahead, the clear next step is to repeat this research using a larger and more diverse sample, ideally drawing lecturers from multiple private university campuses across Colombo district and, where feasible, other regions of Sri Lanka, following the full sampling plan already outlined using the Krejcie and Morgan approach discussed in the methodology. A larger, multi-campus sample would allow the individual contributions of task automation and digital communication tool integration to be tested with far greater statistical confidence, and would make it possible to check whether the construct overlap suspected in this pilot genuinely explains their weaker unique effects, or whether other factors are involved. Future research might also benefit from including lecturers across a wider range of ages and career stages, since the age concentration in this pilot sample may have shaped the results in ways that a more varied group would not replicate. Beyond simply scaling up the sample, later studies could usefully explore some of the qualitative dimensions this quantitative pilot was not designed to capture, such as lecturers' personal experiences of AI training, the specific tools they find most and least helpful, and how institutional culture shapes whether AI is experienced as a genuine resource or as an added source of pressure.

In conclusion, this pilot study provides early but genuinely useful evidence that AI adoption and content creation tools, in particular, can meaningfully support workload management among lecturers in Sri Lankan private universities. While the sample size and context mean that these results should be treated as preliminary rather than conclusive, they offer both a sound theoretical starting point for future research into technology adoption in Sri Lankan higher education and a reasonably clear practical direction for universities considering where to invest their training resources first. With a larger, multi-campus follow-up study, these initial findings could be tested more rigorously and, if confirmed, could meaningfully inform how private universities across Sri Lanka support their academic staff in adapting to an increasingly AI-integrated working environment.

References

Bariham, I. (2026). AI adoption in higher education institutions: A systematic analysis of AI competences, utilisation patterns and their influence on the 21st century skills development in social studies students. Frontiers in Education. Article 1846197. (DOI: 10.3389/feduc.2026.1846197)

Cianci, J., Weibel, D., & Elfering, A. (2024). Measuring work demands and resources of digitalisation: The ICT resources and stressors scale. Swiss Psychology Open, 4(1), 1-26. (DOI: 10.5334/spo.59)

Demuyakor, J., Abudulai, J. I., Baah-Ofori, R., Amankwah, A. S., Segtub, M., Avenyo, S. J., & Boye-Doe, Z. (2026). Unveiling the influence of digital tools for team communication in the digital workspace: Some evidence from three Ghanaian public universities. International Journal of Information and Education Technology, 16(4). (URL: https://www.researchgate.net/publication/403824881)

Ejaz, F., Sultana, S., Rashid, S., Aziz, M. A., Ejaz, S., & Hossain, M. B. (2025). Modelling university lecturers' intention to adopt artificial intelligence in assessment design systems. Innovations in Education and Teaching International. (DOI: 10.1080/14703297.2025.2517295)

Gârdan, I. P., Manu, M. B., Gârdan, D. A., Negoiță, L. D. L., Paștiu, C. A., Ghiță, E., & Zaharia, A. (2025). Adopting AI in education: Optimizing human resource management considering teacher perceptions. Frontiers in Education, 10. Article 1488147. (DOI: 10.3389/feduc.2025.1488147)

Guo, Y., & Wei, L. (2025). AI technology adoption and intergenerational knowledge transfer among older employees. Frontiers in Psychology, 16. Article 1673730. (DOI: 10.3389/fpsyg.2025.1673730)

Kazmi, S., & Irshad, M. (2025). The role of digital communication tools and work-life balance in enhancing employee productivity in hybrid work environments. Contemporary Journal of Social Science Review, 3(3). (URL: https://contemporaryjournal.com/index.php/14/article/view/1143)

Liu, Z., Cao, S., Zhang, J., Chen, Y., Phongsatha, T., & Phongsatha, S. (2026). Understanding university teachers' continuance of an AI teaching assistant: An integrated TTF-TAM-ECM model in higher education. Frontiers in Psychology. Article 1765263. (DOI: 10.3389/fpsyg.2026.1765263)

Pasha, M., Nadeem, S., Rasool, A., Shafique, F., Imran, H., Kanwal, S., Arshad, M., Ali, H., & Yasmeen, R. (2025). Digital fatigue and cognitive overload as managerial challenges: Understanding the impact of workload management on the workforce. The Critical Review of Social Sciences Studies, 3(4), 1629-1641. (DOI: 10.59075/xwtjqy32)

Atherton, T. J., Descamps, I., Holmes, T. R., Vizcarra, C. L., Sui, N., Webel, M., & Foley IV, J. J. (2026). Faculty orientations shape adoption of AI in research and teaching. arXiv preprint arXiv:2605.18140. (URL: https://arxiv.org/abs/2605.18140