Jianli Pan, an Associate Professor of Information Sciences and Technology in George Mason’s College of Engineering and Computing, and Md Arafin, an Assistant Professor of Cyber Security Engineering in the same college, have secured $500,000 from the U.S. National Science Foundation for a project aimed at modernizing Internet of Things and edge cybersecurity research and curricula. The funding began in September 2026 and runs through late August 2029, giving the team a three-year window to build both the science and the teaching infrastructure needed for what they describe as the embodied AI era, a period in which intelligent systems no longer merely analyze data in the cloud but act in the physical world through devices embedded in homes, workplaces, and critical infrastructure.

The central premise of the project is that foundation models, the large-scale neural networks that have transformed fields from language processing to computer vision, are now migrating onto resource-constrained devices at the network edge. This migration creates a fundamentally new security landscape. When a foundation model runs on an edge gateway, a smart camera, or an industrial controller, it must be deployed, optimized, and defended within tight limits on memory, computation, and energy. Those constraints open attack surfaces that conventional cybersecurity curricula rarely address, from model extraction and adversarial manipulation to compromises of the underlying system software that hosts the model. Pan and Arafin intend to study these challenges directly while translating their findings into classroom-ready material.

The research component of the project focuses on emerging foundation models for the edge, examining how such models can be deployed and optimized on IoT platforms without sacrificing security. The investigators plan to develop innovative research and learning materials that address the full lifecycle of edge-deployed intelligence, from initial deployment decisions through performance optimization to the design of protective mechanisms. This dual emphasis on capability and security reflects a growing recognition in the field that AI systems at the edge cannot be treated as ordinary software. Their statistical nature, their dependence on large training datasets, and their integration with physical sensors all create vulnerabilities that demand new analytical tools and new system architectures.

Education sits at the heart of the effort. The researchers will teach emerging foundation-model concepts to students at different educational levels, an unusually broad ambition that spans the pipeline from early coursework to advanced graduate research. To do this, they will create structured modules that give students advanced research and experiential learning opportunities in foundation-model-enabled cyber analytics at the edge. Experiential learning, in which students confront real systems and real constraints rather than abstract exercises, has become a priority for cybersecurity educators who argue that the field’s rapid evolution leaves traditional lecture-based instruction perpetually behind the threats it must counter.

A distinctive element of the project is its attention to secure kernels and system architectures for edge-deployed foundation models. The team will develop innovative training materials on this topic, giving students a grounding in how the lowest layers of an edge system can be hardened to protect the intelligent services running above them. A secure kernel provides a trusted foundation on which higher-level AI workloads can operate, isolating them from tampering and limiting the damage an attacker can do if a device is compromised. Teaching this material alongside foundation-model deployment is intended to produce engineers who understand the entire stack, from silicon-level protections to the behavior of a neural network responding to sensor input.

The project’s outcomes are designed to reach far beyond a single university. Pan and Arafin plan to disseminate their results through a new textbook dedicated to the subject, a resource that could quickly become a reference for institutions racing to add edge AI security to their programs. The dissemination plan also includes hands-on laboratories supported by complementary virtual-machine-based tools, which allow students to experiment with edge deployments and security configurations without needing specialized hardware in every classroom. Virtualized laboratories have proven especially valuable in cybersecurity education because they permit safe, repeatable experimentation with attacks and defenses that would be impractical or irresponsible to stage on live networks.

Training modules and workshops will extend the project’s reach to audiences beyond enrolled students, offering pathways for practitioners and educators to update their skills as foundation models spread through the IoT ecosystem. A cybersecurity competition rounds out the plan, giving participants a high-stakes arena in which to apply what they have learned. Competitions have long served as a proving ground for cybersecurity talent, and anchoring one in the emerging domain of edge AI security could help establish the problem sets and defensive techniques that the next generation of practitioners will need to master.

The strategic significance of the award lies in its timing. Industry analysts project tens of billions of connected devices in the coming years, and the arrival of embodied AI, in which models perceive and act through physical hardware, multiplies both the opportunities and the risks. Securing that ecosystem requires a workforce fluent in two domains that have historically been taught separately: distributed systems and networking on one side, and machine learning on the other. By fusing them into a single research and teaching program, Pan and Arafin are betting that the most effective defense against future attacks on intelligent edge infrastructure will be built in the classroom as much as in the laboratory.

For George Mason University, Virginia’s largest public research university, the project reinforces a broader push into engineering and computing at scale. The institution enrolls more than 40,000 students from 130 countries and all 50 states, and its proximity to Washington, D.C. places it near a dense concentration of federal agencies and defense contractors with acute interests in IoT and edge security. Over the three-year funding period, the award is expected to produce research findings, educational materials, and trained graduates whose influence may extend well past the university’s campus, shaping how an entire generation of engineers designs and defends the intelligent, connected devices that are rapidly becoming inseparable from everyday life.

Subject of Research: NSF-funded modernization of IoT and edge cybersecurity research and education using foundation models

Article Title: Pan & Arafin modernizing IoT/Edge cybersecurity research & curricula

Article References: Pan & Arafin modernizing IoT/Edge cybersecurity research & curricula. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: Internet of Things, edge computing, cybersecurity, foundation models, embodied AI, National Science Foundation, George Mason University, curriculum development, secure kernels, experiential learning, edge AI security, workforce training

The ECHO model was originally developed at the University of New Mexico as a response to a stubborn problem: specialist expertise is concentrated in urban academic centres, while patients, including children with complex conditions, are scattered across vast rural and regional areas. Rather than referring patients away, ECHO inverts the traditional clinic model. Primary care providers join regular virtual sessions with a hub of specialists, presenting real de-identified cases and receiving mentorship. Over time, the goal is that local clinicians acquire the skills and confidence to manage increasingly complex cases themselves, building workforce capacity where it is most needed. The approach has been described as a movement of knowledge rather than a movement of patients, and it has spread from hepatitis C management to dozens of clinical domains worldwide.

While the model has been extensively assessed in adult medicine, its use in paediatrics has been reported inconsistently, making it difficult for health planners to know whether the adult evidence can be extrapolated to children’s services. Sheikh and colleagues set out to close that gap by systematically searching MEDLINE, Embase and Web of Science from database inception to 21 May 2025, with no language restrictions. Eligible studies were empirical evaluations of paediatric ECHO programs that reported implementation outcomes. Two reviewers independently screened records and extracted data, and study quality was assessed using the Mixed Methods Appraisal Tool, a validated instrument for judging the methodological rigour of quantitative, qualitative and mixed methods research.

The review’s central analytical choice was Proctor and colleagues’ outcomes framework, a taxonomy that has become a cornerstone of implementation science. The framework distinguishes early implementation outcomes, such as feasibility, acceptability, adoption and appropriateness, from later outcomes including fidelity, penetration, sustainability and cost. Feasibility asks whether a program can actually be delivered as intended; acceptability captures how satisfied participants are; adoption measures how many settings or providers take the program up; and appropriateness reflects how well the program fits local needs. The later outcomes probe deeper: fidelity examines whether the intervention is delivered as designed, penetration measures how far it reaches within a target system, sustainability asks whether it endures, and cost quantifies the resources consumed.

The results revealed a striking asymmetry. Evidence for the early outcomes was consistently strong. Feasibility was rated as high in 48 of the 55 studies, or 87 percent, acceptability in 33 studies, or 60 percent, adoption in 29 studies, or 53 percent, and appropriateness in every single study, a remarkable 100 percent. In other words, wherever paediatric ECHO programs have been tried, evaluators have found them workable, welcome and well matched to clinical need. For a model that depends on busy clinicians volunteering time to join virtual sessions alongside their clinical duties, that level of buy-in is a meaningful signal that the model resonates with frontline paediatric providers.

The later outcomes told a very different story. Fidelity was examined as high in only five studies, or 9 percent, penetration in 18 studies, or 33 percent, and sustainability in 12 studies, or 22 percent. Cost, arguably the outcome most relevant to health system decision-makers deciding whether to fund expansion, was assessed as high in no studies at all. This means that although the field has accumulated abundant evidence that ECHO can be launched and appreciated in paediatric settings, it knows remarkably little about whether the sessions are delivered faithfully to the model’s core design, whether they reach a meaningful share of the intended workforce, whether they survive beyond initial grant funding, or what they actually cost per clinician trained or per patient reached.

The review did identify patterns that distinguished higher performing programs. Those that succeeded often incorporated structured session formats, interactive case discussions and institutional support. These elements align with the model’s theoretical underpinnings: structured formats keep sessions focused and reproducible, interactive case-based discussion is the engine of knowledge transfer in ECHO, and institutional support addresses the practical barriers of protected time, technology and administrative backing that can otherwise cause programs to wither. For organisations planning to launch paediatric ECHO programs, the implication is that attention to program architecture and organisational buy-in matters as much as the clinical content itself.

Methodological quality, however, varied across the evidence base, and the review was candid about its limitations. Most of the 55 studies used descriptive or mixed methods designs, and the quality assessment identified common weaknesses relating to confounding, sampling and the integration of quantitative and qualitative components. The authors also noted that sex and gender were not consistently reported, an omission with potential relevance to understanding how different groups of clinicians and patients engage with and benefit from the model. These gaps mean that the strong early-outcome findings, while encouraging, rest largely on evaluations that were not designed to withstand rigorous causal scrutiny.

The consequences of this evidence gap are not merely academic. Health systems facing paediatric workforce shortages, particularly in rural and underserved communities, are increasingly looking to tele-mentoring models as a scalable solution. Without data on fidelity, penetration, sustainability and cost, planners cannot distinguish between a program that genuinely builds durable local capacity and one that produces enthusiastic testimonials but little measurable change. The authors argue that embedding standardised assessments of these later outcomes into future programs is essential to support evidence-based planning and scalable implementation. In practice, that would mean building evaluation instruments into program design from the outset, rather than treating evaluation as an afterthought once a program is already running.

The review ultimately delivers a two-sided verdict on one of the most influential healthcare training innovations of the past two decades. On one side, ECHO has demonstrated near-universal feasibility and appropriateness in paediatric care, with strong acceptability and solid adoption, confirming that the model’s blend of virtual specialist mentorship and case-based learning translates effectively to children’s health services. On the other side, the near-total silence on cost, and the sparse evidence on fidelity, penetration and sustainability, leave the most important questions unanswered. As funders and health departments weigh investments in decentralised paediatric care, the message from this review is clear: the model works well enough to warrant expansion, but the field must now prove, with rigorous standardised measurement, that it lasts, spreads and pays off.

Subject of Research: Implementation of the ECHO tele-mentoring model in paediatric healthcare

Article Title: Implementation of the extension for community healthcare outcomes model in paediatric care: a systematic review

Article References: Sheikh, M., Khatri, D., Williamson, K., & Conyers, R. (2026). Implementation of the extension for community healthcare outcomes model in paediatric care: a systematic review. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07790-7

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07790-7

Keywords: ECHO model, paediatrics, tele-mentoring, implementation science, systematic review, decentralised care, workforce training, health services, Proctor outcomes framework, BMC Pediatrics, rural healthcare, capacity building

The stakes are considerable. Australia produced roughly 369,000 tonnes of fresh bananas in the 2023–24 season, worth an estimated $697.8 million, and about 94 per cent of that output comes from Far North Queensland. The industry is built around approximately 540 family farms alongside a smaller number of corporate enterprises, and despite decades of mechanisation it remains fundamentally labour-intensive. Harvesting, de-handing, grading, packing, and ripening all depend on timely, skilled human intervention. When labour is unavailable, mismatched, or insufficiently trained, fruit loses market value quickly, even while remaining nutritionally sound.

What makes the study distinctive is its mesoeconomic lens. Rather than treating food waste as a technical failure at a single point in the supply chain, the researchers examine the intermediate layer of institutions and organisations, such as farms, packing sheds, labour pipelines, and market channels, that connects individual farm decisions to economy-wide outcomes. Food waste, in this framing, is the emergent result of coordination, or miscoordination, between labour supply, production schedules, quality standards, and market access. The study deliberately distinguishes waste generation from value recovery, treating bananas diverted to processing, secondary markets, donations, or animal feed as recovered rather than simply discarded.

The four case studies span the industry’s full structural range. Farm 1 is a small family operation under 50 acres relying almost entirely on household labour, selling directly to local and interstate markets and reporting waste of only about 5 per cent, much of it diverted to animal feed. Farm 2 is a small-to-medium enterprise dependent on rotating visa-based crews of Pacific Island workers and backpackers. Farm 3 is a medium-to-large centralised packing shed serving seven growers with a mixed workforce of permanent staff, long-term migrants, backpackers, and PALM scheme workers. Farm 4 is a vertically integrated giant with more than 500 workers, its own institutionalised PALM pipeline, robotic packing, and processing facilities that recover roughly 60 per cent of its off-spec fruit.

One of the study’s most important findings is that labour shortages rarely translate directly into unharvested bananas. Instead, the labour–waste relationship operates through subtler channels: workforce continuity, farm-specific knowledge, training requirements, and task accuracy. On Farm 2, each new intake of visa workers must be retrained in precise tasks such as bell injection, spraying, bagging, and harvesting, where mistakes cause cosmetic damage that leads to downgrading. Farm 3 reported that its real bottlenecks were not general labour but role-specific positions in maintenance, logistics, and semi-technical work, where instability could disrupt equipment reliability, grading consistency, and transport coordination. Farm 4, by contrast, deliberately engineered workforce stability through its PALM scheme, keeping around 250 workers active from a rotating pool of about 400 trained returnees, so that only about 10 per cent of each intake were new and unskilled.

Skill itself emerged as a system property rather than a fixed attribute of individual workers. On the smallest farm, competence was embedded in lived, intergenerational experience, with the grower explaining that misplaced chemical injections during bell injection could directly damage fruit and create waste. On the larger farms, skill was formalised through induction, licensing, structured toolbox talks, and in-house agronomy, with Farm 4 describing a philosophy of constant retraining so that workers do not create waste in the first place. The progression from tacit know-how to precision-based craft to process-managed competence mirrors the farms’ increasing organisational complexity, and demonstrates that training is simultaneously a workforce-development strategy and a food-waste prevention intervention.

Market structures proved just as decisive as labour. Farm 2’s waste was driven primarily by supermarket cosmetic grading regimes and domestic oversupply rather than any failure of production, with rejected bananas described as nutritionally perfect yet commercially unsaleable. The grower deliberately limited second-grade sales, fearing that cheap downgraded fruit would erode first-grade prices in a market supplied entirely domestically. Farm 3’s packing shed quantified the seasonal rhythm of loss, with waste rising to 25–30 per cent in the summer months and falling to 15–18 per cent in winter, compounded by bruising from transport. Strikingly, the shed found that packing to different retailer specifications changed both waste rates and kilograms packed per worker, a direct illustration of how downstream retail standards propagate upstream into labour productivity and reject volumes.

The study also uncovered a non-linear relationship between technology and waste that challenges assumptions about agricultural modernisation. The smallest farm achieved low waste with almost no technology, relying on rapid market turnover and close supervision. The intermediate farms adopted sensors, drones, and digital monitoring, yet experienced higher waste because deeper integration into supermarket supply chains exposed them to stricter grading and logistics pressures. Only at the highest level of integration, exemplified by Farm 4’s mechanised sheds, robotic pallet stacking, cold-chain infrastructure, and processing plants that convert imperfect bananas into peeled product for banana bread manufacturing and skins into cattle feed, did technology combine with organisational capability to recover substantial value. Even there, recovery decisions were governed by cost structures rather than technical feasibility alone.

The researchers are careful to note the limits of their evidence. The four cases were selected for analytical diversity rather than statistical representativeness, the design is qualitative and cross-sectional, and the perspectives of workers, retailers, and processors were not directly captured. They also caution that the operational benefits of the PALM scheme should not be read as an endorsement of its broader social performance, given documented concerns about worker rights, employment conditions, and dependency on employers under temporary migration arrangements.

Nevertheless, the policy implications are clear. Reducing food waste in perishable horticulture requires integrated interventions that go well beyond farm-level technology adoption: strengthening labour pipelines, supporting skill development, expanding economically viable secondary markets, and investing in processing infrastructure. Interventions should be differentiated by scale, since small farms face affordability constraints while large enterprises face pipeline governance and retention challenges. Above all, the study reframes food waste as a mesoeconomic coordination outcome, produced at the interface between farms and markets, and shows that whether a banana becomes waste depends less on how it was grown than on how the whole system of labour, grading, logistics, and recovery is organised around it.

Subject of Research: Labour demand–supply coordination and food waste generation and recovery in the Australian banana industry

Article Title: From plantation to secondary market: A mesoeconomic thematic analysis on labour demand–supply coordination and food waste reduction in the Australian banana industry

Article References: Roy, M., Akbar, D., Rahman, A., Rajapaksa, D., & Rixon, C. (2026). From plantation to secondary market: A mesoeconomic thematic analysis on labour demand–supply coordination and food waste reduction in the Australian banana industry. Journal of Agriculture and Food Research, 31, Article 103329. https://doi.org/10.1016/j.jafr.2026.103329

Image Credits: AI Generated

DOI: 10.1016/j.jafr.2026.103329

Keywords: food waste, banana industry, labour coordination, mesoeconomics, Queensland agriculture, circular economy, PALM scheme, supermarket grading standards, secondary markets, workforce training, supply chain, value recovery

The timing of the intervention is significant. The commentary forms part of a series of contributions to the 2026 UN Water Conference, supported by the United Nations University Institute for Water, Environment and Health as part of the UN-Water initiative to embed evidence-based policymaking in the water sector. As governments gather to negotiate responses to escalating water crises, Ong and Tortajada make the case that workforce transformation deserves the same political attention as pipes and plants. The authors acknowledge funding support from PUB, Singapore’s national water agency, an organisation frequently cited as a global benchmark for integrated water management and one that has invested heavily in developing specialised talent.

The pressures bearing down on urban water systems are well documented. Climate change is intensifying both floods and droughts, forcing utilities to manage extremes that legacy systems were never designed to withstand. Ageing infrastructure — much of it buried beneath city streets for half a century or more — is failing at rising rates, with pipe bursts, sewer overflows and treatment plant breakdowns imposing growing costs on operators and communities. Urban expansion adds another layer of stress, as informal settlements and fast-growing peri-urban districts often develop faster than formal water networks can reach them. Meanwhile, technological change is rewriting the skills profile of the sector entirely, as digitalisation, automation, remote sensing and artificial intelligence enter control rooms that were once run on manual gauges and experience-based judgement.

Each of these pressures translates into a human challenge. Consider the demographic problem first. Across many developed economies, the water workforce is ageing faster than the broader labour market, and retirements are outpacing recruitment. Studies cited in the commentary, including analysis by Brookings on renewing the water workforce in the United States, describe a widening pipeline gap: fewer young people enter water-related trades and professions than the sector needs to replace those leaving. The United States Environmental Protection Agency has repeatedly highlighted this workforce constraint alongside its funding programmes, including the water infrastructure investments announced under the bipartisan infrastructure law, which explicitly call for prioritising underserved communities. Money, the argument goes, cannot buy competence; somewhere between the appropriation of funds and the delivery of a resilient system stands a technician who must be trained, certified and retained.

The technological dimension of the workforce problem is equally consequential. Research published in Environmental Science & Technology by Garrido-Baserba and colleagues has examined how the digitalisation of wastewater operations changes decision-making in utilities, introducing data-driven control strategies that demand new competencies. The commentary extends this analysis to cybersecurity, an increasingly urgent concern as water and sewer infrastructure becomes networked. Work from Kompetenzzentrum Wasser Berlin on the future evolution of water infrastructure and associated cyber risks, together with industry analyses from GHD on overcoming vulnerabilities in critical water infrastructure, illustrates how digital transformation creates exposure that only a skilled, security-aware workforce can manage. A treatment plant that can be operated remotely can also, in principle, be attacked remotely — and the first line of defence is not software but the human expertise layered around it.

Different jurisdictions are already experimenting with different workforce models, and the commentary draws together these scattered examples into a coherent agenda. In the United States, the EPA maintains dedicated resources on the water infrastructure sector workforce, framing workforce development as integral to infrastructure sustainability. In Australia, the New South Wales Department of Climate Change, Energy, the Environment and Water has published a water industry traineeship guide designed to create structured entry pathways into the sector. New Zealand has gone further with a national competency framework maintained by Water New Zealand, which sets out the skills and standards expected of water professionals across the industry. In the United Kingdom, the sector skills body Energy & Utility Skills published a workforce renewal and skills strategy in 2020 anticipating demographic decline and skills shortages across energy and water utilities, while case studies such as Thames Water’s diversity and inclusion programme demonstrate attempts to widen the recruitment base.

Singapore offers perhaps the most comprehensive illustration of what deliberate workforce transformation looks like. PUB’s long-running innovation series documents how the city-state has paired infrastructure investment with sustained research and development capacity, cultivating in-house expertise in membrane technology, water reuse and integrated catchment management. The lesson, Ong and Tortajada suggest, is that technology adoption and workforce development are not parallel tracks but a single system: advanced treatment processes only deliver value when engineers understand them, operators can run them and managers can plan around them. UNESCO’s 2016 World Water Development Report, subtitled Water and Jobs, made a related macro-level argument nearly a decade ago — that water and employment are mutually reinforcing — but the commentary argues that this insight has still not been translated into routine workforce planning within urban utilities.

The health dimension of the workforce question also deserves emphasis. Research from New Zealand, published in the New Zealand Medical Journal by Chambers and colleagues, has examined drinking-water quality outcomes in ways that ultimately trace back to the competence of those operating supplies. When treatment fails, the public health consequences can be immediate and severe, as outbreaks of waterborne disease in various countries have demonstrated. This elevates workforce capability from a sectoral concern to a matter of public safety, comparable to aviation safety culture or medical training standards. Yet the water sector rarely enjoys the professional visibility of those industries, and the commentary argues that raising the status, pay and career prospects of water workers is therefore not merely an employment issue but a structural investment in public health and trust.

What, concretely, would a transformed urban water workforce look like? Drawing on the evidence assembled in the commentary, several elements emerge. Utilities and governments would treat workforce planning as a core function, mapping future skill needs against climate scenarios and digital roadmaps rather than filling vacancies reactively. Education and training pipelines would be formalised through apprenticeships, traineeships and competency frameworks of the kind already piloted in Australia, New Zealand and the United Kingdom. Recruitment would broaden beyond traditional engineering cohorts to include data scientists, cyber specialists and community engagement professionals, alongside deliberate efforts to diversify the workforce and include underrepresented groups. Retention would be addressed through career progression, fair compensation and professional recognition, countering the attrition that drains experienced operators into other industries. And international cooperation — the spirit of the 2026 UN Water Conference process — would help transfer these models to cities in lower-income countries, where workforce gaps are often most acute and where urbanisation is fastest.

The broader message of the commentary is one of reframing. Water security is usually measured in cubic metres, infrastructure asset registers and investment sums. Ong and Tortajada ask policymakers to add a different metric: does the city have enough people with the right skills to design, build, operate, secure and adapt its water systems under conditions of accelerating change? Their answer, informed by evidence spanning the United States, Europe, Australia, New Zealand and Singapore, is that most cities do not — and that this human deficit is quietly undermining decades of infrastructure investment. If the urban century is to be water-secure, they conclude, the transformation of the urban water workforce is not an optional complement to infrastructure spending but its essential foundation. The pipes, in other words, are only as resilient as the people who run them, and the time to invest in those people is now.

Subject of Research: The role of workforce transformation in supporting urban water security.

Article Title: Transforming the urban water workforce will support water security

Article References: Ong, C., & Tortajada, C. (2026). Transforming the urban water workforce will support water security. Nature Water. https://doi.org/10.1038/s44221-026-00708-8

Image Credits: AI Generated

DOI: 10.1038/s44221-026-00708-8

Keywords: urban water workforce, water security, water infrastructure, climate change, digitalisation, cybersecurity, workforce training, competency frameworks, urban water management, Nature Water, UN Water Conference, water utilities