Knowledge, Innovation and Technology in Scientific Research Vol. 2

Circular Economy Business Models: A Critical Review of Value Creation, Scalability and Economic Impact

Anil Kumar — 2026-09-15

Circular economy business models are widely presented as a means of separating commercial value creation from primary resource throughput, yet the commercial and macroeconomic evidence supporting that proposition remains uneven. This critical narrative review examines three interlocking questions: how circular business models create, deliver and capture value; why comparatively few of them reach a scale at which they alter industrial or consumption patterns; and what can defensibly be concluded about their economic impact at firm and economy-wide levels. Literature published between 2015 and 3 July 2026 was identified through structured searching of open scholarly databases and indexes, supplemented by backward and forward citation searching and by targeted retrieval of institutional evidence, with foundational earlier work retained where conceptually necessary. Sources were appraised for design adequacy, measurement validity, treatment of confounding and external validity rather than by citation count alone. Four conclusions emerge. Conceptual proliferation has outpaced empirical consolidation, since competing typologies describe similar value logics in incompatible vocabulary, which frustrates cumulative measurement. Evidence that circular practices are associated with improved financial performance is reasonably consistent in direction but rests heavily on cross-sectional, self-reported and disclosure-derived measures, so causal interpretation remains weak. Scalability is constrained less by technology than by value capture, capital structure, ecosystem coordination and consumer acceptance, and the small number of studies that examine scaling directly point to capability and governance explanations that have not been tested at population level. Economy-wide assessments report modest and conditional aggregate gains, with employment and emissions outcomes depending on policy design, accounting boundaries and rebound assumptions. The persistent disconnect between encouraging firm-level associations and modest system-level effects is the central unresolved problem in this field. Research priorities include longitudinal and quasi-experimental designs, standardised economic indicators at value-chain level, and explicit treatment of rebound within business model evaluation.

Edge Artificial Intelligence and Ultra-Low-Power Sensor Networks for Closed-Loop Precision Irrigation: A Critical Review of Hardware Architectures, Algorithms and Field Evidence

Bhupendra Dhankar — 2026-09-15

Closed-loop precision irrigation couples continuous measurement of soil, plant or atmospheric state to automatic actuation of water delivery, and the migration of inference from remote servers to battery-powered field devices has been presented as the development that makes such control practical at farm scale. This review critically evaluates the evidence supporting that proposition across three interdependent layers: the sensing and network hardware that constrains what can be measured and transmitted, the algorithms that convert measurements into irrigation decisions under severe memory and energy budgets, and the field deployments that test whether the resulting systems conserve water without agronomic penalty. Literature published between January 2015 and the final search date was identified through open scholarly databases and indexes, supplemented by targeted verification of bibliographic records and by searches of institutional sources, with foundational earlier work retained where conceptually necessary. The synthesis indicates that the strongest and most reproducible evidence concerns component-level performance. Quantised neural networks now execute on microcontroller-class hardware at sub-millijoule energy per inference, low-power wide-area radios support kilometre-scale field coverage, and soil-specific calibration reduces low-cost dielectric sensor error to values approaching those of research-grade instruments. Evidence for the closed loop as a whole is considerably weaker. Reported water savings, which cluster between roughly fifteen and forty-five per cent, derive predominantly from single-season, spatially unreplicated comparisons against loosely specified conventional practice, and the most striking algorithmic gains originate in crop-model simulation rather than in commanded irrigation. Energy autonomy is frequently asserted from datasheet arithmetic that field measurements contradict by an order of magnitude. Persistent gaps include the absence of shared benchmarks and open field datasets, the near-total lack of multi-season controlled trials with hydraulic verification of delivered water, and a geographical concentration of deployments that limits transferability to the smallholder systems most often invoked as beneficiaries. Confidence in current water-saving estimates should therefore remain provisional.

CRISPR and Beyond: Next-Generation Genome Engineering for Sustainable and Climate-Resilient Agriculture

Ashutosh Gautam — 2026-09-15

Agricultural genome engineering has moved rapidly from double-strand-break-mediated mutagenesis towards base editing, prime editing, cis-regulatory engineering, epigenome editing, targeted integration and increasingly tissue-culture-free delivery. This expansion is often presented as a direct route to climate resilience and sustainable intensification, but molecular precision does not by itself establish durable agronomic benefit. This critical narrative review evaluates how next-generation genome-engineering platforms can contribute to crop adaptation to drought, salinity and temperature extremes, disease resistance, nutrient-use efficiency and yield stability, while examining the translational constraints that separate successful editing events from deployable cultivars. Literature published from 2012 to 6 July 2026 was identified through multidisciplinary and agriculture-focused scholarly sources, supplemented by citation searching and bibliographic verification. Evidence was appraised according to editing strategy, molecular validation, genetic background, regeneration and delivery route, phenotyping environment, agronomic endpoints, replication and translational relevance. The evidence is strongest where editing changes endogenous regulation rather than simply disrupting genes, where plausible physiological mechanisms are linked to yield under stress, and where effects have been reproduced beyond controlled environments. Field-validated examples nevertheless remain much less common than greenhouse or growth-chamber demonstrations. Delivery and regeneration remain major genotype-dependent bottlenecks, although developmental regulators, ribonucleoprotein delivery, meristem targeting and viral systems are reducing reliance on stable transgenes and conventional tissue culture. Base and prime editing expand allelic precision, while large-DNA integration and chromosome engineering could enable more ambitious redesign, but their efficiency, predictability and crop generalisability are still uneven. Sustainable outcomes also depend on trait architecture, genotype-by-environment interactions, stewardship, regulation, intellectual-property conditions and access to transformation or editing infrastructure. The next phase of agricultural genome engineering should therefore be judged by field durability, resource and risk outcomes, genetic-background breadth and equitable deployability, rather than editing efficiency alone.

Strategic Agency Reserve in AI-Intensive Organisations: Preserving Human Judgement, Intervention Capacity and Resilience under Intelligent Automation

Kwan Hong TAN — 2026-09-15

Aims: This chapter proposes Strategic Agency Reserve (SAR), an original management construct describing the stock of human capability that allows an organisation to diagnose, judge, intervene, improvise and accept accountability when AI-mediated routines become unreliable, ambiguous, unavailable or misaligned.

Study Design: Integrative conceptual review and theory-development chapter.

Methodology: The analysis synthesises published peer-reviewed research on generative AI productivity, human-automation reliance, organisational learning, skill formation, strategic decision making, organisation design and resilience. Conceptual comparison is used to distinguish SAR from human capital, absorptive capacity, dynamic capabilities, slack and resilience.

Results: The chapter identifies five dimensions of SAR: epistemic vigilance, contextual judgement, intervention competence, adaptive recombination and accountability ownership. It develops ten propositions explaining how AI delegation can erode or replenish SAR, proposes a 15-item diagnostic scale and complementary behavioural indicators, and specifies reserve-preserving practices for AI-intensive organisations. The framework predicts that high delegation can remain strategically robust when organisations deliberately maintain a sufficient and distributed reserve, whereas high delegation combined with low reserve creates brittle autonomy.

Conclusion: AI strategy should treat human capability not simply as labour awaiting substitution, but as an option-bearing reserve whose value becomes most visible during exceptions, shocks and model failure. SAR offers a measurable research agenda and a practical design principle for combining AI-enabled efficiency with durable organisational agency.