The distinction matters because a framework can define what professionals should know and be able to do without ensuring that they can perform those tasks in a real clinical environment. Capability describes the required knowledge, skills and behaviours. Competence concerns whether an individual has attained those capabilities at an appropriate standard, while performance is their demonstrated application in practice. Workforce readiness goes further: it includes the judgement, confidence, ethical reasoning and adaptability needed to use those capabilities in complex settings that change as technologies, regulations and patient needs evolve. The perspective also uses Miller’s pyramid as an interpretive lens, with readiness reflected in the ability to translate knowledge and competence into effective performance in authentic practice settings.

Australia’s coordinated policy work is described as a significant achievement. The National Digital Health Workforce and Education Roadmap established the case for a nationally consistent approach, while the National Nursing and Midwifery Digital Health Capability Framework, the Digital Health in Medicine Capability Framework and the Digital Health Capability Framework for Allied Health Professionals translated the broader policy direction into professional contexts. The cross-disciplinary Australian Digital Health Capability Framework then articulated shared expectations across the health and care workforce, including digital professionalism, leadership and advocacy, and the quality of data and information. Although the frameworks address different professional settings, the authors identify substantial common ground across digital professionalism, information and data literacy, technology-enabled care, human-centred practice and safe use of digital technologies.

These initiatives have helped move digital health beyond a narrow measure of technical proficiency. Their shared language covers not only the operation of digital systems, but also communication, human-centred care, collaboration, safe technology use and responsible handling of information. Australia has also begun building implementation support, notably through the Digital Health Train-the-Trainer Toolkit, developed by the Australian Digital Health Agency with the Australian Council of Senior Academic Leaders in Digital Health. It provides resources for university educators embedding eight digital health learning outcomes in pre-registration health degrees, recognising that educators themselves need support if curricula are to change. The initiative shifts attention from further refining capability statements toward the practical work of curriculum integration, educator preparation and implementation support.

The perspective does not present framework development as a completed solution. Many competency requirements have yet to be systematically embedded in health professional curricula, and evidence for sustained improvements in organisational and workforce outcomes remains limited. Reviews have found variation in the design of digital readiness interventions, their implementation and the measures used to assess success. Educational programs can strengthen digital competence and confidence, but the authors say that the evidence for long-term effects remains less developed. Their proposed response is not to discard capability frameworks, which will need to evolve with technology and regulation, but to connect existing expectations to training, assessment, simulation, workplace learning and continuing professional development. The success of the national agenda would then be judged by what becomes observable in education and practice rather than by the number of frameworks produced.

Authentic practice learning is central to that proposed transition. Students and health professionals must encounter situations in which they must judge whether information is reliable, protect privacy, navigate ethical dilemmas, support consumers with different levels of digital literacy and respond when a technology or process fails. A readiness indicator under the framework’s data and information quality domain, for example, would be the ability to critically evaluate data and AI-generated outputs before using them in clinical decisions. In technology domains, readiness would be demonstrated by recognising failures, seeking support and activating safe alternatives to maintain continuity and care. Under digital professionalism, the proposed indicator is the application of professional judgement when technologies create inequities, uncertainty or ethical dilemmas involving access, privacy, security, communication and AI-related issues. These are illustrative transformations of framework statements, not outcomes measured in the perspective.

Artificial intelligence makes this distinction increasingly consequential. The source argues that workforce preparation must extend beyond conventional technical competence to include assessment of AI-generated content, awareness of automation bias and hallucinated outputs, and recognition of the limitations of algorithmic recommendations. These are not simply additional software skills: they require professional judgement, data stewardship, privacy awareness, accountability, bias awareness and the ability to question an apparently authoritative output. The emerging position presented in the paper is that AI should augment rather than replace professional judgement, and that responsible innovation requires ethical governance, human-centred design and workforce leadership. Readiness therefore means being prepared for digitally mediated decision-making, not merely having access to a tool or knowing how to operate it.

Readiness also depends on the organisations in which professionals work. Leadership support, reliable digital infrastructure, interoperability, governance, workflow integration, technical assistance, workplace culture and protected learning opportunities all affect whether a person can translate training into everyday practice. The perspective therefore treats readiness as both an individual attribute and a product of a supportive organisational environment. Equity is an equally important concern because infrastructure, access to professional development and previous experience with digital tools may differ across metropolitan, regional, rural and remote settings, professional groups and career stages. Exposure to digital technology or access to a course consequently cannot be treated as evidence of readiness by itself. Planning must account for a person’s goals, motivations, literacy, aptitude and context as well as their formal access to training.

The proposed next phase calls for sustained, coordinated implementation and evaluation across universities, accreditation bodies, health services, professional organisations, policymakers and researchers. Questions include whether digital capabilities are consistently taught and validly assessed, whether educators are equipped to support learning, whether simulation and workplace placements provide authentic experience, and whether graduates and practising clinicians continue to demonstrate safe, responsible and adaptive digital practice over time. The Australian Council of Senior Academic Leaders in Digital Health supports a broader systems approach involving education, workforce development and digital health research, while DHWorx, an Australian and New Zealand digital health workforce research consortium, provides a mechanism for collaborative scholarship about capability, preparedness and emerging educational needs. Their value will depend on evidence showing that implementation is sustained rather than remaining aspirational.

The perspective is primarily Australian and draws on selected national initiatives and literature rather than a systematic review of international evidence. It also identifies substantial limitations in the empirical record: long-term effects of digital capability frameworks on the workforce remain limited, and readiness requirements will differ among professions, career stages, organisations and healthcare settings. Rapid technological change further means frameworks and readiness measures will need regular revision. The central message is nevertheless clear: Australia has made major progress in defining digitally capable practice, but producing a ready workforce now depends on measured implementation, supported learning, organisational conditions and continuous evaluation rather than the publication of more capability documents alone. The proposed transition is therefore from defining what digitally capable practice should look like to testing whether health professionals can enact it safely, responsibly and adaptably as healthcare changes.

Subject of Research: Australian digital health workforce capability and readiness

Article Title: Beyond capability frameworks: Australia’s next challenge in preparing the digital health workforce

Article References: Arabi, S., & Butler-Henderson, K. (2026). Beyond capability frameworks: Australia’s next challenge in preparing the digital health workforce. Frontiers in Health Services, 6, Article 1968499. https://doi.org/10.3389/frhs.2026.1968499

Image Credits: AI Generated

DOI: 10.3389/frhs.2026.1968499

Keywords: digital health, workforce readiness, capability frameworks, health education, artificial intelligence, digital literacy, organisational readiness, professional development, health policy, Beyond, capability, frameworks

One learner admitted, "I use Alexa every day but didn't know it was AI until you told me." Another described using ChatGPT to generate bedtime stories for a son without ever thinking of the chatbot as artificial intelligence. The adaptation reflects a growing recognition in the literature that for populations facing structural and social obstacles, technology acceptance is not purely a cognitive calculation but a socially mediated process shaped by instructors, peers, and institutional conditions. The analysis followed Moustakas' phenomenological procedure.

The researcher first practiced epoche, deliberately bracketing personal assumptions about AI in education to minimize bias, then applied horizontalization, treating every significant statement from the transcripts as equally important before condensing them into meaning units and clustering them into themes. From this process, four themes crystallized: understanding AI in everyday life, building confidence and a growth mindset around technology, personalized empowerment through AI, and willingness to learn amid structural barriers. Trustworthiness was reinforced through member checking, in which participants reviewed their transcripts, along with reflexive journaling to document analytic decisions.

The second theme revealed how confidence actually forms. Participants described initial hesitation rooted in limited digital literacy and fear of making mistakes. "I was scared of it at first. But once I used it a few times, I realized it is not that bad," one learner reflected.

Another captured the social dimension of ease of use: "If someone shows me how to use it, I feel better. On my own, I get frustrated." This finding challenges the assumption that ease of use is a property of software design alone. For adult learners, perceived ease of use was socially constructed, built through instructor guidance, peer support, and low-stakes opportunities to experiment.

Confidence, in other words, is not downloaded; it is scaffolded. Once confidence took root, AI transformed from a novelty into a personalized learning partner. Participants praised tools like Grammarly and Google Translate for offering judgment-free practice, a benefit that proved especially powerful for English language learners. "I like that I can use Grammarly or Google Translate and not feel embarrassed if I make a mistake," one participant said.

Another compared ChatGPT to a tutor who adjusts to the learner: "ChatGPT explains things in a way I understand. It's like having a tutor who adjusts to me." The tools also extended learning beyond classroom hours. "It helps when I'm working at night or early in the morning. I do not have to wait for help," one learner explained, underscoring how AI's always-on availability fits the unpredictable schedules of adults juggling employment, caregiving, and coursework.

Yet enthusiasm collided with structural reality. Others felt overwhelmed by the pace of technological change layered on top of foundational computer skills they were still acquiring. Privacy concerns and uncertainty about how the systems work added further hesitation.

The result was a persistent gap between willingness and opportunity: learners consistently expressed a desire to engage with AI if someone would simply show them where to start. Without structured instruction, the study suggests, AI integration risks reinforcing existing inequities rather than closing them, particularly for learners from lower socioeconomic backgrounds or those with interrupted educational histories. The findings also carry implications for workforce preparedness, though the author is careful about the causal chain.

The study did not measure direct employment outcomes; instead, it found that AI use supported foundational academic skills, writing proficiency, and digital confidence, competencies that modern workplaces increasingly demand. Michel frames this as a developmental pathway: AI first builds educational readiness, which then becomes the foundation for workforce participation in technology-driven labor markets. Within the adapted TAM framework, trust emerged as a prerequisite for perceived usefulness, meaning learners were unlikely to view AI tools as beneficial at all unless they first felt confident in their accuracy and fairness.

The study's limitations are explicit. Eight participants drawn from a single region of Southern California cannot statistically represent the world's adult learners, and recruitment through educator referrals may have selected for learners already somewhat interested in technology. The snapshot design also captures perceptions at a single moment, early in what may be a long evolution of awareness and trust.

Subject of Research: Adult learners' perceptions, trust, and barriers regarding artificial intelligence use in adult education

Article Title: Investigating adult learners’ perceptions of artificial intelligence in relation to knowledge, trust, learning outcomes, and workforce preparedness

Article References: Michel, F. (2026). Investigating adult learners’ perceptions of artificial intelligence in relation to knowledge, trust, learning outcomes, and workforce preparedness. Discover Education, 5(1), Article 913. https://doi.org/10.1007/s44217-026-02134-0

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02134-0

Keywords: artificial intelligence, adult education, technology acceptance model, digital literacy, AI trust, workforce readiness, ESL learners, qualitative research, educational technology, lifelong learning, AI literacy, Southern California

Labour literacy is a concept that extends well beyond vocational training. The researchers define it as the full set of work-related knowledge, skills and attitudes that students need to develop both personally and professionally. It encompasses an understanding of labour markets and workers’ rights, practical competencies for employment, and the dispositions, such as responsibility and perseverance, that underpin sustainable working lives. Unlike narrow technical skills, labour literacy is conceived as a foundational component of a student’s overall development, which is precisely why educators and policymakers have struggled to measure it, let alone cultivate it at scale.

The research team, led by Pengfei Liu under the supervision of Piyapong Sumettikoon, surveyed 472 university students, including both undergraduate and postgraduate learners, all of whom had prior experience with digital learning. Participants completed structured questionnaires measuring their perceptions of digital platform support, their learning motivation, their self-regulated learning habits, and their labour literacy. The responses were analyzed in SPSS using descriptive statistics, reliability and validity testing, correlation and regression analysis, and structural equation modeling, a statistical technique that allows researchers to test whether one variable influences another indirectly through an intermediary.

The headline result is a statistically significant positive association between digital platform support and labour literacy, with a standardized path coefficient of 0.350, significant at the p less than 0.001 level. In practical terms, students who reported that their platforms provided useful resources, timely feedback and responsive technical support also reported higher levels of work-related knowledge, skills and attitudes. But the more intriguing findings lie in the mediating pathways. Digital platform support was associated with higher learning motivation, which in turn was linked to labour literacy, with a mediated path coefficient of 0.277. A second pathway ran through self-regulated learning, the capacity to plan, monitor and adjust one’s own studying, which carried a coefficient of 0.250.

The structural equation models revealed striking explanatory power. The full model accounted for an R-squared value of 0.791 for labour literacy, meaning that the combination of platform support, motivation and self-regulated learning statistically explained nearly 79 percent of the variance in students’ labour literacy scores. Learning motivation itself showed an R-squared of 0.839, and self-regulated learning an R-squared of 0.815, suggesting that the variables in the model capture a tightly interwoven web of educational dynamics rather than isolated factors.

These results align with well-established theoretical frameworks. Self-determination theory, developed by psychologists Edward Deci and Richard Ryan, holds that motivation flourishes when learners experience autonomy, competence and relatedness, conditions that well-designed digital platforms can actively supply through choice, immediate feedback and collaborative tools. Meanwhile, Bandura’s social cognitive theory emphasizes that learners develop self-efficacy by observing models, mastering experiences and receiving constructive feedback, all of which digital environments can deliver at scale. The new findings suggest that when platforms satisfy these psychological needs, they do more than deliver content; they cultivate the motivational and regulatory machinery that students carry into their working lives.

The study’s practical implications are considerable. For universities, the message is that investment in digital infrastructure should not be measured solely by content delivery or user satisfaction metrics, but by whether platforms are designed to foster motivation and self-regulatory skill. Features such as adaptive learning pathways, progress dashboards, goal-setting tools and structured feedback loops are not merely conveniences; according to this model, they are the mechanisms through which technology translates into workforce readiness. The authors recommend that educational institutions and government agencies integrate digital platforms deliberately into labour training programs rather than treating them as supplementary tools.

For policymakers, the findings arrive at a moment when governments are grappling with how to close skills gaps identified in reports from organizations such as QS and Cengage Group, which document persistent mismatches between graduate capabilities and employer expectations. If labour literacy can be strengthened through the motivational and self-regulatory affordances of digital platforms, then national digital education strategies may double as workforce development strategies, a proposition with significant budgetary and curricular implications. The researchers suggest that technology-enhanced labour education could be scaled in ways that traditional classroom-based career preparation cannot easily match.

As with any cross-sectional survey study, caution is warranted in interpreting the direction of causality. The data capture associations at a single point in time, and it remains possible that students who are already more motivated and self-regulated engage more deeply with digital platforms, rather than platforms producing those qualities. Longitudinal and experimental designs would be needed to confirm the causal pathways the model proposes. Nevertheless, the strength of the observed coefficients, the large sample size and the use of structural equation modeling give the findings considerable weight, and the authors note that the anonymized dataset has been deposited in a public Zenodo repository to allow independent verification and reuse.

What makes this study resonate beyond the statistics is its reframing of a familiar debate. Rather than asking whether screens help or harm learning in the abstract, it identifies the specific psychological levers, motivation and self-regulation, through which digital tools appear to shape one of the most consequential outcomes in higher education: whether students leave university genuinely prepared for the world of work. In an era when the half-life of job-specific skills is shrinking and lifelong learning has become an economic necessity, understanding how everyday educational technology builds the deep literacy of labour may prove to be one of the most consequential questions educational science can ask.

Subject of Research: How digital learning platforms enhance university students' labour literacy through learning motivation and self-regulated learning

Article Title: Enhancing students’ labour literacy through digital platform support: the mediating role of learning motivation and self-regulated learning

Article References: Liu, P., & Sumettikoon, P. (2026). Enhancing students’ labour literacy through digital platform support: the mediating role of learning motivation and self-regulated learning. SN Social Sciences, 6(9), Article 385. https://doi.org/10.1007/s43545-026-01661-9

Image Credits: AI Generated

DOI: 10.1007/s43545-026-01661-9

Keywords: labour literacy, digital platforms, learning motivation, self-regulated learning, educational technology, structural equation modeling, higher education, employability, technology-enhanced learning, student engagement, workforce readiness, survey research

The paper, authored by Christina Neider, Ed.D., Vice Provost of Colleges, and Marc Booker, Ph.D., Vice Provost for Strategy, describes the university’s institutional framework for embedding generative AI across academic programs, student learning experiences, and the processes, policies, and workflows that keep a large online institution running. Rather than treating AI as a single tool to be bolted onto existing systems, the authors present a three-pillar structure intended to balance innovation, ethical use, and learner outcomes. The framework is explicitly aligned with the needs of working adult learners, a population the university serves almost exclusively, and with the evolving expectations of employers who increasingly want graduates who can use AI effectively and responsibly.

The first pillar focuses on embedding AI into programs and course content. The authors argue that learners now need three interlocking competencies: foundational knowledge of how generative AI works, practical skills for applying it in professional contexts, and the critical judgment required to evaluate AI-generated output. To build these competencies, the university has made a generative AI platform available to all students and offers an elective introductory course on generative AI. A set of brief, self-paced AI essentials modules has been embedded in the learning management system, allowing students across every program and course to develop AI literacy and confidence at their own pace. The Phoenix Success Series, the introductory course sequence taken by first-year students, will introduce generative AI as a support tool for learning beginning in 2026.

The scale of the curriculum work described in the paper is notable. The university has deployed an interactive tool for creating Socratic dialogues in five courses, reaching approximately 1,300 students across disciplines through scenario-based practice that is reviewed by faculty. This approach reflects a deliberate pedagogical choice: rather than letting students passively consume AI output, the Socratic format requires them to engage in structured dialogue, defend reasoning, and refine understanding, with human instructors retaining oversight of the exercise. In parallel, more than 20 degree programs have been strategically selected for curriculum revision, with at least two AI-integrated summative assessments planned in each program by March 2027. Summative assessments, which evaluate what students have learned at the end of a course, represent a high-stakes point of integration, signaling that AI competence is being treated as a graduation-level outcome rather than an optional enrichment.

The second pillar addresses the use of AI tools to enhance the learning experience itself. Here the emphasis shifts from what is taught to how students and faculty navigate their academic environment. The paper describes a range of use cases, some fully available and others in pilot stages, spanning every level of the institution. Students have access to general AI assistance, while academic services questions can be answered within the classroom environment itself. Faculty receive their own AI assistance tailored to their instructional needs. Perhaps most significantly, course-specific AI assistants support students with content and curriculum inquiries directly connected to the courses in which they are enrolled, providing a layer of just-in-time academic help that sits between broad student services and individual faculty office hours.

This tiered support architecture is one of the more technically interesting aspects of the framework. By connecting AI assistants to specific course content, the university reduces the risk of generic or inaccurate responses and grounds the assistance in the actual materials students are studying. The distinction between general assistance, service-oriented support, and course-connected help illustrates how AI deployment in education benefits from deliberate layering rather than a single monolithic chatbot. It also reflects a broader principle in human-centered AI design: tools should meet users where they are, in the context of their immediate task, rather than requiring them to seek help elsewhere.

The third pillar extends AI integration beyond students to faculty, staff, policies, training, and operational practices. The authors emphasize that meaningful implementation cannot stop at the classroom door. The university is applying AI to improve both academic and business efficiencies, and the paper argues that ethical AI use requires institutions to translate policy into practice. A philosophy on generative AI established in 2023 has informed subsequent policy development, faculty guidance, and the creation of AI literacy resources designed to support responsible use across the institution. The university’s Center for AI Resources, described in the paper as foundational to this effort, offers resources for students, faculty, and staff and has reached more than 90,000 users to date.

The framework’s intellectual grounding is the Digital Education Council AI Literacy Framework, a human-centered model that the authors cite as informing the university’s commitment to foundational AI knowledge, practical skills, and responsible-use habits. This alignment matters because it situates the university’s approach within an emerging international consensus about what AI literacy means. Literacy, in this view, is not merely technical fluency. It encompasses the judgment to know when AI output should be verified, the ethical awareness to recognize bias and misuse, and the habit of treating AI as a collaborator whose contributions must be critically assessed. Dr. Booker frames the institutional imperative in similarly holistic terms, noting that the rapid evolution of generative AI is forcing institutions to transform and that the answer lies in integrating thoughtfully across governance, tools, processes, and curriculum, with policy strengthening cross-departmental connections as well as operational and academic functions.

The workforce dimension runs throughout the paper and gives the framework much of its urgency. As generative AI reshapes work across industries, employers increasingly expect candidates who can use these tools productively while understanding their limitations. Dr. Neider argues that working adult learners need meaningful opportunities to practice, improve outcomes, and understand the ethical considerations that accompany AI use, and that AI literacy requires human judgment as well as practice. By embedding AI throughout the academic experience rather than confining it to a single course, the university aims to help students develop the skills and responsible-use habits that employers increasingly value. For adult learners who are often studying while working, the immediate applicability of these skills is a practical benefit as much as an academic one.

What emerges from the white paper is less a prescription than a proof of concept: a demonstration that AI integration can be organized, measured, and scaled when treated as an institution-wide undertaking rather than a collection of disconnected experiments. The three pillars, curriculum, learning experience, and institutional process, are designed to reinforce one another, so that policy informs practice, practice informs pedagogy, and pedagogy serves the ultimate goal of student success. For a sector still wrestling with questions of academic integrity, equity of access, and the pace of technological change, the paper offers a concrete data point in an ongoing experiment. Whether the model transfers to institutions with different missions and resources remains an open question, but the underlying logic, that AI adoption must be holistic, human-centered, and tied to real workforce demands, is likely to shape the conversation well beyond a single university.

Subject of Research: Institutional framework for integrating generative artificial intelligence in higher education

Article Title: Three-pillar framework for integrating AI in higher education outlined in University of Phoenix white paper

Article References: Three-pillar framework for integrating AI in higher education outlined in University of Phoenix white paper. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: generative AI, higher education, AI literacy, University of Phoenix, curriculum design, working adult learners, educational technology, AI ethics, institutional policy, workforce readiness, Digital Education Council, white paper

Researchers led by Nishat Tabassum of Pabna University of Science and Technology, working with colleagues from Comilla University, surveyed students at three major public institutions: the University of Dhaka, Rajshahi University, and Pabna University of Science and Technology. Using stratified random sampling, they collected responses from 150 students at each university, drawn evenly from engineering, science, social sciences, arts, and humanities faculties. The respondents, 52.2 percent male and 47.8 percent female and mostly between 18 and 25 years old, answered a structured questionnaire administered through a web survey, with the resulting data analyzed using IBM SPSS Statistics 20 and Excel.

The intellectual scaffolding of the study rests on three theories. Technological determinism, rooted in the work of sociologist Thorstein Veblen, holds that technological innovation drives social change and reshapes institutions, often faster than those institutions can adapt. Skill mismatch theory describes the gap that emerges when workers’ competencies diverge from labor market demands, whether through horizontal mismatch between education and job requirements, vertical mismatch of overqualification, or skills obsolescence caused by rapid technological turnover. Labor market theory, meanwhile, frames how shifts in the supply and demand for particular skills translate into changes in wages and employment levels. Together, the researchers argue, these frameworks explain why students expect AI to transform their careers and why they fear the education system is not keeping pace.

The headline finding is cautiously optimistic: across all three universities, a majority of students believe AI will generate more employment opportunities than it eliminates in their fields. Rajshahi University students expressed the greatest optimism, with 55.3 percent agreeing that AI will expand job opportunities, followed by PUST at 52.7 percent and Dhaka University at 43.3 percent. This pattern echoes a growing international literature suggesting that AI’s effect on employment is better characterized as job shifting than mass job loss, as automation displaces routine tasks while creating new occupations that demand advanced cognitive, digital, and socio-emotional capabilities.

Beneath that optimism, however, lies a deep vein of anxiety and institutional criticism. More than half of the students at every university surveyed agreed that AI-driven change would produce a skills mismatch, with agreement peaking at 62 percent at Rajshahi University. Even more telling, 70.7 percent of all respondents believed a gap exists between the skills they are currently learning and the skills required for AI-related jobs. And when asked whether their institutions were actively updating curricula to include AI-relevant skills, 62.7 percent disagreed overall, with dissatisfaction reaching 70 percent at Dhaka University and 69.3 percent at PUST. Only at Rajshahi University did a comparatively larger share, 37.3 percent, believe their curriculum was being meaningfully refreshed.

The study also documented institutional variation in how much AI education students are actually receiving. Rajshahi University reported the best integration, with 48.7 percent of students saying they receive adequate AI education in their programs, against 38.6 percent who disagreed. At Dhaka University the picture was more divided, with 48.6 percent disagreeing that their AI education was sufficient, and at PUST a majority, 50.6 percent, said their programs were not preparing them adequately. Correspondingly, only 12.7 percent of Dhaka students and 14.7 percent of PUST students rated the availability of AI-related courses and training at their institutions positively, compared with 24 percent at Rajshahi University. The researchers note that AI course integration appears greatest in engineering faculties and weakest in the sciences across all three institutions.

Gender and geography shaped perceptions in subtle ways. Female students at Dhaka University and PUST reported higher familiarity with AI than their male peers, at 75.6 percent and 88.4 percent agreement respectively on the relevance of AI to their fields, while at Rajshahi University males reported greater familiarity and a sizable share of female respondents, 19.4 percent, remained neutral, a pattern the authors suggest may reflect unequal access to AI-focused learning opportunities. Regional dynamics were equally revealing: Dhaka University students, situated in the capital with greater industry exposure, were the least fearful of job loss, with 56.7 percent disagreeing that AI would cause significant job losses, while roughly half of PUST and Rajshahi students expressed such fears, a disparity the researchers attribute to uneven access to technological infrastructure and employment networks outside Dhaka.

Where students agreed almost unanimously was on the enduring value of human skills. Majorities at all three institutions, ranging from 66.7 percent at Rajshahi to 73.3 percent at PUST, said jobs requiring creativity and emotional intelligence are less likely to be automated. Even more striking, between 73.3 and 84 percent of students at the three universities said AI knowledge is key to future career success. Yet this awareness did not translate into confidence: 52.7 percent of Dhaka University students took a neutral stance on their own ability to adapt to AI-driven change, compared with far lower neutrality and higher confidence at PUST, where 50.7 percent agreed they could adapt, and Rajshahi, where 55.3 percent did so.

The demand signal from students could hardly be clearer. Around 81 percent of PUST students and 82 percent of Rajshahi students, along with 72 percent of Dhaka students, called for more practical AI training in their curricula, and similar majorities wanted more hands-on AI experience and stronger collaboration between universities and industry. Confidence in government action was notably weaker: a net negative assessment of policy effectiveness emerged, with 44 percent of respondents disagreeing that government initiatives were addressing AI-era skills mismatches, against just 30 percent agreeing. This skepticism persists despite Bangladesh’s National Strategy for Artificial Intelligence, launched in 2020, and a National AI Policy adopted in 2024, suggesting a perceived gap between policy formulation and implementation on the ground.

The authors conclude that without deliberate intervention, the promise of AI-driven development in Bangladesh may remain unrealized. They recommend embedding foundation AI courses across all faculties, not just engineering; organizing workshops, seminars, and guest lectures with AI professionals to expose students to real-world applications; aligning university curricula with national AI policies; investing in faculty training and the infrastructure needed for AI teaching and research; and fostering continuous learning habits among students so their skills remain relevant throughout their careers. For a country where nearly 34 percent of young people are considered tech-savvy and the youth population is expanding rapidly, the stakes are considerable. The study suggests that Bangladesh’s students have already made up their minds about AI: they see it as an opportunity, but one their universities have not yet equipped them to seize.

Subject of Research: University students' perceptions of artificial intelligence's impact on job prospects and skill mismatch in Bangladesh

Article Title: Understanding university students’ perspective on artificial intelligence and its influence on job prospects and skills in Bangladesh

Article References: Tabassum, N., Ferdous, J., Hossain, M. S., & Haque, T. (2026). Understanding university students’ perspective on artificial intelligence and its influence on job prospects and skills in Bangladesh. Discover Global Society, 4(1), Article 237. https://doi.org/10.1007/s44282-026-00413-7

Image Credits: AI Generated

DOI: 10.1007/s44282-026-00413-7

Keywords: artificial intelligence, Bangladesh, university students, job prospects, skill mismatch, higher education, curriculum reform, labor market, AI literacy, employment, public universities, workforce readiness