A socio-technical mechanism of the agricultural digital divide: Ecuador within a heterogeneous Latin American context

1 Introduction

Digital technologies are entering agricultural production, farm management, extension services, finance, marketing, and supply chains. Precision agriculture, digital platforms, Internet of Things applications, remote sensing, and artificial intelligence can expand available information and support more timely decision-making. However, the existence of these technologies does not imply that all agricultural actors participate in digital transformation under equivalent conditions. Recent literature documents differences associated with costs, competencies, farm characteristics, infrastructure, technical support, financing, institutional arrangements, and technological design (Chesi et al., 2026; Damasceno et al., 2026; Lei et al., 2026; Puntel et al., 2023).

This distinction matters because digital agriculture and the agricultural digital divide describe different phenomena. Digital agriculture refers to the set of technologies, data, services, and digitally mediated processes applied to agriculture of which precision agriculture constitutes one domain. The agricultural digital divide, by contrast, refers to inequalities among agricultural actors in their capacity to participate in digitalization and to benefit from it. Low adoption of precision technologies may be a manifestation of inequality, but it does not by itself define the divide.

It is also necessary to distinguish access from outcomes. Connectivity or device availability can create an opportunity for participation, but they do not guarantee the capabilities needed to use agricultural tools effectively. Capabilities, in turn, do not guarantee adoption, and adoption does not necessarily imply sustained or contextualized use. Evidence on smallholder farmers shows that affordability, skills, institutional support, content relevance, language, and technological suitability can condition use even when basic connectivity exists (Gatti et al., 2026; Mollel et al., 2025; Omwenga et al., 2026). Lei et al. (2026), in a meta-analysis of 49 studies and 125,998 participant observations in China, find that digital literacy is a relevant determinant of adoption and that basic Internet infrastructure is not necessarily the main limiting factor once a minimum level of connectivity already exists.

Economic consequences require equal precision. Digital technologies can improve efficiency without producing equivalent changes in income or wellbeing, while farm scale, financing, business models, platform dependence, and bargaining conditions can affect both the viability of investment and the distribution of the value generated by digitalization (Carolan, 2020; Florez et al., 2026; Hackfort, 2023; Rosnan and Yusof, 2023). The relevant question, therefore, is not only who accesses or adopts technology, but who can appropriate it, integrate it productively, bear its costs, realize economic value, and capture a significant share of the benefits generated.

This article proposes an integrated socio-technical interpretation of these inequalities. Rather than organizing the agricultural digital divide as a list of parallel barriers, it puts forward a mechanism that links structural conditions, resource accessibility, capability formation, technological appropriation, productive integration, economic value realization, and value capture and benefit distribution. The individual relationships draw on existing empirical and conceptual evidence; their integration into a single mechanism constitutes the central conceptual contribution of this perspective article. Ecuador is used as a focal analytical case and is contrasted with Latin American and international evidence, without assuming that it is statistically representative of the region.

2 Evidence base and analytical approach

This perspective article is based on a targeted bibliographic and documentary evidence base, not a systematic review. The analytical corpus was formed from bibliographic records previously compiled by the authors and subsequently consolidated, screened for relevance, and deduplicated by DOI and title. The literature spans 2019–2026 and covers digital inequality, adoption of agricultural technologies, socio-technical systems, digital agriculture, smallholder farmers, agricultural economics, platforms and data governance, artificial intelligence, and international experiences relevant to the proposed framework.

Evidence selection followed a targeted approach guided by the article's conceptual question. Priority was given to studies on digital inequality, digital agriculture, capabilities, technological appropriation, and economic outcomes. Research on data governance, institutional conditions, and comparative agricultural experiences was also considered. Records were screened for thematic relevance, duplicates were verified by title and DOI, and technological studies with no clear connection to the socio-technical and economic mechanisms under analysis were excluded. The corpus made it possible to examine the relationships among structural conditions, resources, capabilities, appropriation, and productive integration, and was likewise used to analyze economic value realization, value capture, and benefit distribution. For Ecuador, priority was given to official statistical sources with verifiable indicators, population, time period, and geographic scope. The resulting framework constitutes an analytical synthesis grounded in the evidence reviewed and requires subsequent empirical validation.

3 From digital access to agricultural digital inequality

The agricultural digital divide should not be confused with digital agriculture or with precision agriculture. Digital agriculture comprises technologies, data, and services applied to agricultural activities, whereas precision agriculture constitutes a specific domain within that set. The agricultural digital divide, by contrast, describes inequalities among actors in their access to digital resources and their capacity to use them and benefit from them (Table 1).

Table 1. Translating agricultural digital-divide mechanisms into policy instruments.

The instruments in the table are policy options derived from the proposed mechanism; they should not be interpreted as universally effective interventions. Their effectiveness depends on agrarian structure, institutional capacity, market conditions, technology characteristics, and implementation quality.

In this article, we propose a definition of the agricultural digital divide as the inequality among agricultural actors in their capacity to access relevant digital resources, develop the capabilities required to use them, appropriate and integrate digital technologies into productive activities, and convert that use into economically valuable outcomes under differing institutional, organizational, territorial, and market conditions. This inequality also encompasses differences in the capacity to capture and retain the value generated through agricultural digitalization.

The literature on digital inequality has evolved from access-centered approaches toward multilevel perspectives that distinguish access, use and capabilities, and outcomes derived from digital use (Anaraky et al., 2025; Dipendra et al., 2025; Wen et al., 2023). Building on this, we propose transferring this logic to the agricultural context and extending it through two economic dimensions: value realization and the capacity to capture and retain it. This extension is consistent with evidence showing that digital use does not by itself guarantee higher income or wellbeing, and that relationships with platforms or providers can condition the capture of the value generated (Carolan, 2020; Hackfort, 2023; Prause et al., 2021; Rosnan and Yusof, 2023).

This distinction avoids treating connectivity, adoption, and productivity as interchangeable indicators of the same divide. An inequality in access does not necessarily imply an inequality in adoption, nor does the latter equate to an inequality in outcomes. The mechanism developed in the next section organizes these differences as analytically related transitions, without assuming that they constitute a universal causal chain.

4 A socio-technical mechanism of the agricultural digital divide

The agricultural digital divide can be understood as a cumulative and conditioned process, rather than as a set of independent barriers. Figure 1 summarizes the proposed mechanism. Its components and several of its relationships are supported by the corpus, but the complete sequence should be interpreted as a conceptual synthesis of the present article, not as a universal causal chain already validated in a single study.

4.1 From structural conditions to capabilities

Agricultural actors enter digitalization with different resources and constraints. Farm scale, financial resources, education, territorial location, market position, and access to institutions can condition the digital options available to them. Studies on smallholder farmers consistently identify affordability, connectivity, devices, financing, and institutional support as relevant conditions for digital participation (Chesi et al., 2026; Yuan and Sun, 2024). Access, however, does not automatically produce capabilities. Lei et al. (2026) identify digital literacy as an important determinant of adoption, while other studies highlight training, extension, language, and contextual knowledge. The initial mechanism is therefore more precise than a direct infrastructure–adoption relationship: structural conditions affect resource accessibility, and these resources interact with training, assistance, and experience to form capabilities.

4.2 From adoption to appropriation and productive integration

A farmer may install an application, register on a platform, or acquire a sensor without incorporating it stably into their decisions. Digital services can remain underused when the content is difficult to understand, the language is not appropriate, the design does not respond to the context, or the technology does not fit productive practices (Sarku et al., 2025).

In this article, appropriation is used to distinguish initial adoption from meaningful, contextualized use; it is presented as an analytical tool to describe the shift from having or trying a technology to integrating it into farm practices. Productive integration occurs when that use genuinely influences production, management, marketing, financing, or advisory activities. Extension workers, producer organizations, cooperatives, and technology providers can shape this transition, as they influence interpretation, support, problem-solving, and adaptation (Gardezi et al., 2026; Nettle et al., 2025).

4.3 From productive integration to economic value realization

The economic relevance of digitalization begins when productive integration changes outcomes; these may include cost reduction, greater efficiency, productivity, market access, profitability, or resilience. However, these outcomes do not translate directly into benefit. Rosnan and Yusof (2023) show that the adoption of digital technologies can improve efficiency without necessarily translating into higher income or wellbeing. Xu et al. (2026) find relationships among digital financial inclusion, climate-smart agriculture adoption, supply chain agility, human capital, and profitability. Lyu et al. (2025) further document heterogeneity in benefits according to income level.

Costs are part of the mechanism. Fixed investment, subscriptions, connectivity, training, maintenance, advisory services, switching costs, and data requirements can affect net returns. Farm scale can matter when fixed costs must be amortized over small volumes. Credit can resolve a liquidity problem, but it does not render a technology profitable when its expected benefit is lower than its total cost. This is where the relevance of alternative provision models arises, such as leasing, pay-per-use, shared services, or cooperative provision, whose usefulness depends on the type of technology, local demand, organizational capacity, and governance.

4.4 From value creation to value capture and benefit distribution

Even when digitalization creates economic value, farmers do not necessarily capture all of that value. Digital systems involve platforms, manufacturers, service providers, advisors, and intermediaries whose business models influence the distribution of costs and benefits. The literature on platforms and corporate digitalization has documented risks of technological dependence, data control, and asymmetries between farmers and providers (Carolan, 2020; Hackfort, 2023; Hackfort et al., 2024; Prause et al., 2021). Value creation and value capture must therefore be treated as distinct concepts.

Data governance constitutes a cross-cutting mechanism. Access rights, control, permitted uses, portability, interoperability, and the ability to move information between providers can affect technological dependence and bargaining conditions (Atik and Martens, 2021; Chamorro-Padial et al., 2026; Lara Ruíz and Gallardo Cobos, 2026). Accordingly, we propose a cross-cutting pathway that links data generation, access and control, portability and interoperability, dependence or autonomy, bargaining position, and value capture.

Cooperatives and producer organizations can modify several points in the mechanism. Latha et al. (2025) show that the digitalization of dairy cooperatives can support organizational processes, supply chain management, and performance, while Lei et al. (2026) find a growing influence of cooperatives and similar organizations on adoption. However, their effect should be understood as conditioned by governance, scale, capabilities, and contractual arrangements, not as a universal solution.

5 Ecuador within a heterogeneous Latin America

Ecuador is useful for examining the agricultural digital divide because it allows consideration, within a single national context, of agrarian structure, territorial inequalities, connectivity, financing, and localized technology adoption. Its role in this article is analytical, not representative.

Official statistics show why it is necessary to differentiate among indicators. The Third National Agricultural Census recorded 842,882 agricultural production units (UPAs) in 2000. According to the National Institute of Statistics and Census (INEC) and the Ministry of Agriculture and Livestock (MAG) (2002), approximately 29.5% of UPAs were smaller than one hectare and around 63.5% were smaller than five hectares. INEC (2025) reported that 71.3% of households had Internet access and that 80.1% of people aged five and older had used the Internet during the previous twelve months. These are population-level indicators of access and use, not of agricultural digitalization.

Agriculture-specific evidence is more localized. Peña-Holguín et al. (2025) analyze the adoption of IoT technologies among farmers in Guayas and provide evidence on factors associated with adoption. This study helps identify mechanisms, but it does not allow a national adoption rate to be inferred. Ecuador should therefore be used to observe interactions among productive factors without extrapolating results to the rest of Latin America. Brazil presents an advanced agricultural ecosystem alongside inequalities between agribusiness and family farming (Coca et al., 2025). Adoption there is also linked to scale, credit, costs, and farm characteristics (Damasceno et al., 2026; Ivale et al., 2024). In Chile, Orellana et al. (2025) document precision agriculture among smallholder farmers. Puntel et al. (2023), comparing Brazil, Argentina, Uruguay, and Chile, note that digital agriculture metrics are not standardized, limiting the validity of rankings. Rather than ranking countries using heterogeneous indicators, the comparison is organized around four common mechanisms: governance arrangements, technology-provision models, organizational intermediation, and mechanisms of value creation, capture, and distribution.

International evidence supports this: in the United States, factors such as profitability, trust, costs, and advisor preferences affect technology adoption and use despite the country’s technological advancement (Alkhwaldi et al., 2026; Greig et al., 2024; Zia et al., 2026). In Europe and the United States, although data-governance frameworks exist, these rules do not guarantee an equitable distribution of value (Atik and Martens, 2021; Lara Ruíz and Gallardo Cobos, 2026). In Africa, mobile access improves opportunities, but constraints on affordability, capabilities, extension, language, and contextual design still persist (Chesi et al., 2026; Mollel et al., 2025; Omwenga et al., 2026; Sarku et al., 2025). The key is to understand which institutional and economic mechanisms affect the transition from access to productive use and value capture.

6 From barriers to policy mechanisms

A mechanism-based interpretation transforms public policy. Connectivity, training, credit, and regulation are key, but each intervention must be linked to specific constraints in the socio-technical process. When individual ownership is unviable, shared services or leasing are better than subsidizing purchases. Credit is useful when liquidity problems exist. Training should be linked to concrete agricultural tasks and evaluated through sustained use.

Internet of Things (IoT) devices are oriented mainly toward generating and transmitting data, while AI processes that data for classification, prediction, recommendation, or optimization tasks applied to agricultural decisions. Data quality, local adaptation, costs, human capabilities, and governance determine whether a recommendation can be used safely and productively (Gardezi et al., 2026; Zia et al., 2026).

We propose an AI-readiness filter comprising five questions: Is there a clearly defined agricultural decision that AI can support? Are the data adequate and locally relevant? Is the application economically viable? Can users or advisors interpret its results? And are data rights, use, accountability, and governance sufficiently defined? This filter is a synthesis derived from the mechanisms identified, rather than a previously validated scale.

7 Discussion and research agenda

The main implication of this perspective is that the agricultural digital divide should not be measured solely by technological access, since inequality can emerge at several stages, such as obtaining resources, developing capabilities, integrating technologies, realizing economic benefits, and retaining a share of the value generated.

Evaluating success also changes. A program may improve connectivity without improving agricultural capabilities or ensuring sustainable use; a technology may raise productivity yet still offer weak net returns once costs are considered; and digitalization may create value that ends up distributed among providers or platforms rather than farmers. Each outcome requires its own indicators.

Agricultural economics is central here, since fixed and recurring costs, complementary investments, scale, financing, transaction costs, provision models, profitability, and bargaining relationships affect the sustainability of digitalization—particularly for smallholder farmers, whose returns vary with scale, costs, mode of provision, and capacity to capture the value generated.

Data governance also matters, since farmers are not only users but data creators, and their capacity to manage information can shape dependence and bargaining conditions. Future research should move beyond debates about "data ownership" toward specific rights, contracts, interoperability, provider switching, and economic consequences.

The proposed framework has limitations. It integrates mechanisms documented across different technologies, countries, and research traditions, but has not been validated as a complete causal sequence in a single context; the feedback loops linking retained value to future investment capacity, or low returns to reduced willingness to invest, remain conceptual propositions requiring further study. Ecuador also reveals an information gap: national statistics are reliable but limited regarding the relationship between farm characteristics and digital capabilities, technological appropriation, productive use, economic returns, and data-governance practices. Developing such indicators would allow the agricultural digital divide to be assessed without relying solely on general Internet indicators or local adoption studies.

The research agenda should focus on the transitions that still require validation. Longitudinal, farm-level studies could assess whether capability formation predicts sustained productive appropriation, under what conditions digital use improves profitability and not only productivity, how returns vary by scale and income, and whether cooperative or shared-service models reduce costs and strengthen bargaining power. Evidence is also needed on AI readiness, local model validation, agricultural data governance, and the distribution of value generated by platforms. This agenda does not introduce an object distinct from that of the original manuscript; it delimits the empirical evidence needed to evaluate the proposed mechanism.

The policy implication is precise: an inclusive agricultural digital transition cannot be evaluated solely by the number of farms connected or technologies deployed. Its economic and social importance depends on whether agricultural actors can develop the capabilities to use digital resources meaningfully, integrate them into viable productive activities, and retain enough value for digitalization to be sustainable. In this sense, the agricultural digital divide is best understood as a socio-technical and economic inequality in the conditions for meaningful participation in digitalization and in the benefits that such participation produces.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Tianshu Quan, Nanjing Forestry University, China

Reviewed by: Valerio Di Stefano, Roma Tre University, Italy

Mingsong Sun, Huaqiao University, China