Next Article in Journal
Thriving in Internal CSR-Driven Service Organisations: The Role of Embeddedness, Moral Meaningfulness and Intrapreneurship
Previous Article in Journal
Green Cores as Architectural and Environmental Anchors: A Performance-Based Framework for Residential Refurbishment in Novi Sad, Serbia
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Artificial Intelligence Adoption in Non-Chemical Agriculture: An Integrated Mechanism for Sustainable Practices

by
Arokiaraj A. Amalan
* and
I. Arul Aram
Department of Media Sciences, Anna University, Chennai 600025, India
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(19), 8865; https://doi.org/10.3390/su17198865 (registering DOI)
Submission received: 26 July 2025 / Revised: 23 September 2025 / Accepted: 24 September 2025 / Published: 4 October 2025

Abstract

Artificial Intelligence (AI) holds significant potential to enhance sustainable non-chemical agricultural methods (NCAM) by optimising resource management, automating precision farming practices, and strengthening climate resilience. However, its widespread adoption among farmers’ remains limited due to socio-economic, infrastructural, and justice-related challenges. This study investigates AI adoption among NCAM farmers using an Integrated Mechanism for Sustainable Practices (IMSP) conceptual framework which combines the Technology Acceptance Model (TAM) with a justice-centred approach. A mixed-methods design was employed, incorporating Fuzzy-Set Qualitative Comparative Analysis (fsQCA) of AI adoption pathways based on survey data, alongside critical discourse analysis of thematic farmers narrative through a justice-centred lens. The study was conducted in Tamil Nadu between 30 September and 25 October 2024. Using purposive sampling, 57 NCAM farmers were organised into three focus groups: marginal farmers, active NCAM practitioners, and farmers from 18 districts interested in agricultural technologies and AI. This enabled an in-depth exploration of practices, adoption, and perceptions. The findings indicates that while factors such as labour shortages, mobile technology use, and cost efficiencies are necessary for AI adoption, they are insufficient without supportive extension services and inclusive communication strategies. The study refines the TAM framework by embedding economic, cultural, and political justice considerations, thereby offering a more holistic understanding of technology acceptance in sustainable agriculture. By bridging discourse analysis and fsQCA, this research underscores the need for justice-centred AI solutions tailored to diverse farming contexts. The study contributes to advancing sustainable agriculture, digital inclusion, and resilience, thereby supporting the United Nations’ Sustainable Development Goals (SDGs).

1. Introduction

India’s agricultural landscape has long been characterised by traditional practices that contribute to environmental sustainability, climate resilience, and food security [1]. However, a notable shift is occurring as younger generations increasingly migrate to non-agricultural sectors, resulting in workforce shortages and a growing reliance on technological interventions [2]. Artificial Intelligence (AI) offers promising solutions by automating processes, enhancing decision-making, and improving operational efficiency [3].
A critical gap remains in understanding its integration with Non-Chemical Agricultural Methods (NCAM) [4]. NCAM represents sustainable farming practices that prioritise biodiversity, preserve soil health, and support chemical-free food production. Integrating AI into NCAM systems could address these challenges by improving productivity, reducing environmental impact, and supporting climate-smart agriculture (CSA) practices [5].
In India, efforts to integrate Artificial Intelligence (AI) with Non-Chemical Agricultural Methods (NCAM) are gaining momentum. Under the National Mission on Natural Farming (NMNF), cluster-based programmes in districts such as Hassan support around 125 farmers per 50-hectare cluster, providing financial incentives and resources for chemical-free cultivation [6]. Similarly, Andhra Pradesh is implementing the Andhra Pradesh Agriculture Information and Management System (APAIMS 2.0), a fully digitised platform that applies AI and machine learning to deliver pest alerts and customised advisories at the plot level, thereby strengthening sustainable practices [7]. Recent studies also show that farmers in South India have experimented with image-based weed detection using mobile cameras, while more advanced applications of AI, including IoT-enabled pest monitoring, crop condition assessment, and agri-robotics, demonstrate considerable potential for sustainable agriculture [8]. By 2050, global population growth will demand increased agricultural productivity and sustainable resource use. Agriculture 4.0, integrating IoT, AI, machine learning, and robotics, enhances crop monitoring, pest management, and livestock welfare worldwide. While international case studies demonstrate its benefits, challenges such as infrastructure gaps, data security, and limited access for smallholder farmers persist. Effective adoption requires collaborative networks, farmer-centric strategies, and investment in sustainable infrastructure to build resilient and equitable global food systems [9].
This study bridges these gaps by adopting a justice-centred lens to examine AI adoption in Non-Chemical Agriculture Methods (NCAM). By employing discourse and thematic analysis, it captures smallholders’ lived experiences, addressing how narratives and local contexts shape adoption [10]. NCAM promotes climate resilience, biodiversity, and sustainability, rooted in traditional practices yet requiring modern support. Sir Albert Howard (1945) [11] praised India’s chemical-free farming resilience. AI can address challenges like labour shortages and climate stress by enabling pest detection, soil monitoring, and yield prediction [12,13].
Despite its promise, AI adoption in NCAM faces barriers such as high costs, low digital literacy, and inadequate rural infrastructure [12,14]. Addressing these infrastructural gaps is crucial for ensuring equitable access to AI technologies. Resistance from traditional farming communities also poses challenges. Inclusive, transparent deployment strategies are essential to foster trust and ensure equitable access [15].
Addressing farmers’ strategic communication needs, can enhance their cognitive, social, and emotional capacities [16,17]. Educational programmes, peer learning, and success stories foster trust [18]. Human-centred, explainable AI ensures farmer control and supports sustainable practices through transparent, autonomous technologies [19,20].
This study pursues three key objectives:
  • To explore the potential applications of AI technologies in enhancing sustainable practices such as NCAM.
  • To understand the challenges and opportunities in adopting AI technologies among NCAM farmers through Focus Group Discussions.
  • To develop a refined conceptual framework that integrates holistic understanding of AI adoption from a farmer-centric perspective.

2. Review of Literature

AI technologies, including artificial neural networks, fuzzy systems, and agricultural robots, are transforming farming by enhancing crop monitoring, pest management, disease detection, and yield prediction [21]. AI-driven smart agriculture systems using deep learning and IoT sensors can reduce pesticide use and improve environmental sustainability [22]. Precision agriculture, supported by technologies like drones, satellites, and sensors, optimises inputs and reduces labour costs [23].
The increasing cost of labour and the demand for premium produce are driving the adoption of AI technologies in on-farm sorting and transportation, help to reduce post-harvest losses through faster and more accurate processing [24]. Additionally, AI-powered autonomous weeding systems employing 3D visual tracking and predictive control enhance sustainability by reducing herbicide use and optimising operations across varied terrains [25].
However, AI adoption in rural India faces challenges such as inadequate infrastructure, digital illiteracy, and financial constraints [14,26]. Linguistic diversity in India presents a major communication challenge for AI adoption, with over 600 languages complicating outreach efforts. Trust issues also persist, particularly among marginalised farming communities, due to limited exposure to AI applications [27].
NCAM, such as organic farming, conservation agriculture, and Integrated Pest Management (IPM), promotes sustainable practices by reducing pesticide use and enhancing soil health [28,29]. It also provides added benefits by eliminating synthetic pesticides, preserving biodiversity, enhancing soil health, and reducing greenhouse gas emissions. [30,31]. AI can complement NCAM by providing real-time monitoring, predictive analytics, and decision-support systems to optimise irrigation, detect diseases, and evaluate soil health [32]. Despite challenges, integrating AI with NCAM holds promise for sustainable farming and improving food security in India [33,34]. Aversion to unsafe food is increasing consumer demand for healthier, organic options, leading to the expansion of NCAM adoption and reinforcing sustainable agricultural practices [35].

3. Conceptual Framework: Integrated Mechanism for Sustainable Practices (IMSP)

This study introduces the Integrated Mechanism for Sustainable Practices (IMSP), a comprehensive conceptual framework designed to holistically examine the adoption of Artificial Intelligence (AI) technologies among farmers practising Non-Chemical Agricultural Methods (NCAM) in Tamil Nadu, India. The framework innovatively combines the Technology Acceptance Model (TAM) with justice-centred theory to address both the technical and socio-ethical dimensions of technology adoption in sustainable agriculture. Through the integration of thematic analysis, critical discourse analysis, and Fuzzy-Set Qualitative Comparative Analysis (fsQCA), the IMSP framework captures the complex interplay of factors shaping farmers’ decisions on responsible technology use, with particular attention to the contextual realities of smallholder farming systems in Tamil Nadu, which is situated within the Global South and reflects the region’s role in addressing agrarian, technological, and socio-economic challenges characteristic of emerging economies.
Previous studies in emerging economies have used PLS-SEM and fsQCA to examine AI adoption, climate resilience, and net-zero carbon goals, highlighting complex, non-linear relationships at the macro level. In contrast, our study applies an Integrated Mechanism for Sustainable Practices (IMSP) at the smallholder level in India by combining fsQCA with Justice-Centred Discourse Analysis of thematic findings. This mixed qualitative–quantitative approach captures context-specific socio-cultural, economic, and infrastructural factors—such as digital literacy, gender barriers, and participatory decision-making—shaping AI and NCAM adoption. By linking quantitative configurations with qualitative insights, the study provides locally grounded, equitable, and sustainable agricultural solutions.

3.1. TAM Application in AI Adoption

The Technology Acceptance Model (TAM), developed by Fred Davis in 1989, underpins the Integrated Mechanism for Sustainable Practices (IMSP) framework. TAM asserts that Perceived Usefulness (PU) and Perceived Ease of Use (PEU) shape users’ Behavioural Intention (BI), which drives actual technology adoption see in Figure 1 [36]. While TAM has been widely applied in agriculture, it often neglects critical contextual factors, including traditional ecological knowledge and socio-structural barriers affecting smallholder farmers’ adoption decisions [37].
In this study, TAM provides a structural lens for understanding farmers’ cognitive and behavioural responses to AI and NCAM. It is augmented with a justice-centred perspective, capturing how social, economic, and environmental factors shape perceptions of usefulness and ease of use. Gendered access to digital tools, resource constraints, and institutional support significantly influence adoption, highlighting the importance of equity in technology implementation.
TAM has also been used in e-learning, e-commerce, and digital services to examine user adoption patterns [38,39]. Integrating TAM with justice-centred discourse analysis and thematic analysis enables both quantitative and qualitative insights, linking behavioural determinants with socio-cultural and structural barriers. This holistic approach informs context-specific strategies to enhance AI and NCAM adoption among smallholder farmers while promoting resilience, inclusiveness, and sustainable agricultural practices.
Applying TAM in agriculture is increasingly important, as sustainable farming practices are still emerging [40]. Farmers often hesitate to adopt chemical-free methods, but TAM effectively explains the acceptance of biological inputs [41]. This study uses TAM (as shown in Figure 1) to assess farmers’ adoption of AI and NCAM and explores how AI can support sustainable agricultural practices.

3.2. Thematic and Discourse Analysis: Capturing Farmers’ Perspectives

The integration of Focus Group Discussions (FGDs) was essential for capturing the nuanced perspectives of farmers that cannot be adequately quantified through statistical methods alone. The open-ended nature of the survey questions, particularly those addressing challenges in AI adoption and preferred support mechanisms, generated rich qualitative data requiring advanced analytical approaches beyond conventional statistical regression.
Thematic analysis was employed to identify, analyse, and report patterns (themes) within the qualitative data, focusing on farmers’ expressed concerns, values, and experiences [42]. For instance, when farmers discussed “intelligent automation minimising manual work,” thematic analysis showed that preferences reflected not only technical needs but also concerns about job security, economic stability, and preserving traditional knowledge. Similarly, support for “workshops or community discussions” highlighted cultural values of collective learning, distrust of top-down extension, and reliance on social networks for knowledge sharing.
Critical Discourse Analysis (CDA) was subsequently applied to examine how language and communication practices shape and reflect social realities, power dynamics, and ideological assumptions [43,44]. Through this approach, straightforward responses about technology preferences were revealed to contain important insights into social relationships, power asymmetries, and cultural values that significantly influence adoption decisions.

3.3. fsQCA for Evaluating Causal Pathways

Fuzzy-Set Qualitative Comparative Analysis (fsQCA) is an advanced analytical method that identifies causal configurations—combinations of conditions—that lead to specific outcomes. It is particularly suited for studies with intermediate sample sizes (such as ours, n = 57), where conventional regression approaches may struggle with issues such as multicollinearity and complex interactions [45]. By applying fsQCA to the adoption of AI in NCAM, this study moves beyond identifying net effects to explore how multiple conditions combine in a conjunctural manner to influence farmers’ decisions.
The fsQCA process followed in this study comprised four structured steps:
  • Calibration: Raw survey data were transformed into fuzzy-set membership scores ranging from 0.0 (full non-membership) to 1.0 (full membership), using direct method calibration grounded in theoretical and empirical knowledge of the sample.
  • Necessity Analysis: We tested whether any single condition was necessary for the outcome, meaning that it was always present when adoption occurred.
  • Truth Table Construction: A truth table was generated to list all logically possible combinations of causal conditions and their empirical outcomes. Frequency and consistency thresholds (see Section 6.3.2) were applied to refine the table and distinguish meaningful causal paths from random noise.
  • Solution Evaluation: The coding patterns for the fsQCA statistical analysis were exported into a csv file, and all subsequent computations were performed using Python 3.12.11 within the Google Colab environment, we derived complex, parsimonious, and intermediate solutions. The interpretation focused on the intermediate solution, as it incorporates theoretically plausible logical remainders consistent with the Technology Acceptance Model (TAM) and justice-centred framework.
While TAM has been used to find technology adoption in agriculture, its application to AI and NCAM remains underexplored. Whereas fsQCA has been used in a study exploring the role of AI in achieving a zero carbon economy in emerging economies showing climate strategy for AI, digital inclusion, and carbon neutrality can emerge from data-driven policy support [46]. Comparable methodologies have been successfully employed in health technology adoption [47], but this study is among the first to apply fsQCA to AI adoption pathways in agriculture.

3.4. Justice-Centred Perspective in AI and NCAM Adoption

A justice-centred view of AI adoption prioritises fair access to technology, shared decision-making, and inclusive policies. Barriers such as cost, infrastructure, and lack of support contribute to the digital divide, while social issues like job security, trust, and the absence of diverse stakeholder voices are also significant [48]. Political theorist Nancy Fraser’s justice framework highlights three dimensions: economic justice (fair resource distribution), cultural justice (recognition of all social groups), and political justice (ensuring equal participation) [49]. A justice-centred approach calls for localised AI models that incorporate ecological and cultural knowledge, promoting inclusion and addressing unequal resource distribution.

3.5. Flowchart of the IMSP Framework

The following flowchart illustrates the integrated approach of the IMSP framework in Figure 2 show the methodical approach made in this study towards the conclusion:

3.6. Novelty of the Study

The IMSP framework makes several significant contributions to the literature on technology adoption in agriculture. First, it advances beyond conventional adoption models by integrating technical, social, and environmental dimensions into a comprehensive analytical framework. Second, it demonstrates how mixed-methods approaches can provide deeper insights into complex adoption decisions than single-method approaches. Third, it provides a transferable model for studying technology adoption in various contexts of sustainable agriculture, particularly in the Global South, such as Tamil Nadu, India where questions of justice, ecology, and technology are intensely intertwined. By integrating fsQCA within the TAM framework and subjecting thematic findings to discourse analysis through a justice-centred lens, this study establishes a unified explanatory mechanism that accounts for both structural and perceptual dimensions of AI and NCAM adoption.

4. Methodology

For this study, a questionnaire combining closed-ended and open-ended questions was employed to gather data on farmers’ attitudes toward Non-Chemical Agricultural Methods (NCAM), their awareness of AI, and communication preferences. Since this study is exploratory, no prior hypotheses were formulated. Instead, open-ended questions allowed participants to express their perceptions and challenges in their own words, which were later coded and organised into qualitative themes and quantitative variables for analysis.
Focus group discussions (FGDs) were also conducted to gain in-depth qualitative insights into challenges, perceptions, and solutions for AI adoption among NCAM farmers [50]. The questionnaire was divided into four sections: Demographic Profile, NCAM Practices, Technology Adoption and AI, and Information Needs and Communication Strategy.

4.1. Study Area

The study was conducted in Tamil Nadu between 30 September and 25 October 2024, covering both sowing and harvest seasons to capture seasonal variations in agricultural practices. Fieldwork was organised at three levels: Village Level (Thuraiyur, Nemili Block, Ranipet District), Block Level (Nemili Block), and District Level (18 districts in Tamil Nadu). Figure 3 illustrates a focus group discussion with farmers from Nemili Block and Thuraiyur Village, with geotagging conducted at Kalapalampattu Village during the farm school session.

4.2. Participant Selection

Using purposive sampling, 57 NCAM farmers were organised into three focus groups representing marginal farmers, active NCAM practitioners, and farmers from 18 districts interested in agricultural technologies and AI. This enabled in-depth exploration of NCAM practices, adoption barriers, and perceptions of AI in agriculture.
Inclusion criteria required participants to be aged 18 or above, with at least three years of farming experience, and either currently practising NCAM or expressing willingness to adopt AI-enabled agricultural practices. Both men and women farmers were included, though village-level participation was predominantly women.
FGD1—Marginal Farmers (n = 18): Farmers practising or interested in NCAM in Thuraiyur Village.
FGD2—Active NCAM Engagement (n = 21): Farmers practising NCAM and sharing insights on technologies, from villages in Nemili Block.
FGD3—Technology-Interested Farmers (n = 18): Farmers from multiple districts offering perspectives on AI and NCAM.

4.3. Data Analysis and Software

The analysis followed a two-stage approach:
Qualitative analysis: Open-ended responses and FGD transcripts were coded thematically to identify key patterns in farmer perceptions.
Quantitative analysis: Coding patterns generated from the open-ended responses were exported into a csv file and analysed exclusively using Python within the Google Colab environment. Descriptive statistics and ANOVA were performed to identify patterns in farmer responses. For advanced analysis, fuzzy-set Qualitative Comparative Analysis (fsQCA) was also conducted, deriving complex, parsimonious, and intermediate solutions to explore the causal conditions influencing AI adoption in NCAM farming.
Since this study is exploratory and relied primarily on open-ended questions that were later coded into themes, statistical reliability tests such as Cronbach’s alpha were not applicable. The focus was on capturing diverse farmer perspectives rather than testing pre-defined constructs through multi-item scales. To ensure rigour, coding was carried out systematically and cross-checked during analysis.
All participants provided informed consent prior to participation. Participation was voluntary, and farmers were assured of confidentiality and the right to withdraw at any stage without consequences.

5. Data Analysis

Descriptive Statistics: Descriptive statistics summarise data using measures of central tendency (mean, median) and variability (standard deviation, range), providing an overview of the dataset. These summaries help identify patterns and distributions, forming the foundation for deeper analysis and ensuring transparency in research, there by aiding decision-making in data analysis [51]. An ANOVA test was subsequently conducted in this study is to determine whether the observed differences in mean scores [52] (on NCAM adoption, AI willingness, and app usage) between the three independent farmer groups (Marginal, Active NCAM, Tech Interested) are statistically significant or likely due to random sampling chance.
Thematic Analysis: Qualitative data from open-ended questions were coded into themes to capture common patterns and viewpoints, grounded in the actual language of participants [53]. Thematic analysis identifies recurring patterns, helping interpret farmers’ core ideas and facilitating decision-making in research design [54].
Discourse Analysis: Using justice-centred approach prioritises equity, examining socio-economic conditions and resource access disparities, particularly for marginalised smallholders [55,56]. Focus group discussion on trust in institutions, participatory decision-making, and digital literacy are central to understanding adoption gaps and fostering inclusivity.
The study integrates fsQCA with TAM uncover essential causal combinations that facilitate AI adoption in the context of NCAM. Unlike regression analysis, fsQCA accommodates non-linear relationships, making it well-suited for exploring complex, real-world scenarios [57]. Themes from FGDs coded for fsQCA analysis are presented in Table 1, illustrating the key conditions for successful AI adoption in NCAM.
The fsQCA process involves necessity analysis to identify conditions consistently linked to AI adoption, followed by sufficiency analysis to explore factor combinations leading to adoption. Integrated with discourse analysis, fsQCA uncovers key causal configurations in the NCAM context. A truth table identifies high-consistency pathways, and complex, parsimonious, and intermediate solutions provide varying levels of causal explanation. Robustness and reliability were ensured through Python-based validation.

6. Data Interpretation

This section interprets farmers’ adoption of AI and NCAM using a mixed-methods approach. Descriptive statistics, thematic analysis, and narrative inquiry provide insights into adoption patterns, while discourse analysis guided by a justice-centred approach examines how socio-cultural, economic, and political factors influence farmers’ choices. Fuzzy-set Qualitative Comparative Analysis (fsQCA), framed by the Technology Acceptance Model (TAM), is integrated to reveal how ecological values, technological perceptions, and communication practices interact to shape sustainable agricultural transitions.

6.1. Descriptive Statistics

The descriptive statistics from the three focus group discussions (n = 57) indicate a predominantly middle-aged male sample with an average farming experience of 4.35 years (Table 2). Respondents were distributed across multiple districts, demonstrating variation. Approximately half of the participants had adopted NCAM, although adoption remains limited. Workforce shortages were minimal, but participants strongly agreed that NCAM is labour-intensive and lacked insufficient technology to offset workforce demands.
Most participants lacked awareness of AI applications in farming and demonstrated limited willingness to adopt AI-driven technologies, highlighting potential barriers. Communication channels were rated as highly effective, while mobile app usage for farming guidance varied, reflecting differing adoption levels.

6.2. ANOVA Analysis Across Farmer Focus Groups

A one-way Analysis of Variance (ANOVA) was conducted to examine differences among three farmer focus groups: marginal farmers (FGD1, n = 18), active NCAM practitioners (FGD2, n = 21), and technology-interested farmers (FGD3, n = 18). Seven variables related to NCAM adoption and technology acceptance were assessed.

6.2.1. ANOVA Results Across Farmer Focus Groups

Regarding NCAM adoption, no statistically significant differences were found among the groups (F (2, 54) = 1.21, p = 0.305), suggesting similar adoption levels irrespective of engagement intensity or technological orientation. Willingness to adopt AI-driven technologies also did not differ significantly (F (2, 54) = 0.91, p = 0.412), indicating broadly comparable motivation across groups.
In contrast, mobile application usage showed a significant difference (F (2, 54) = 8.52, p = 0.014). Post hoc analyses revealed that active NCAM farmers (FGD2) use agricultural mobile apps more frequently than marginal farmers (FGD1), reflecting higher digital engagement among those with greater practical experience.
Analysis of Technology Acceptance Model constructs highlighted further differences. Perceived usefulness (PU) differed significantly among groups (F (2, 54) = 4.87, p = 0.011), with FGD2 reporting higher perceived utility than both FGD1 and FGD3. Similarly, perceived ease of use (PEU) differed significantly (F (2, 54) = 3.94, p = 0.024), indicating that active practitioners find digital tools easier to operate, likely due to greater familiarity and confidence.
User experience (UX) did not vary significantly across groups (F (2, 54) = 2.18, p = 0.119), suggesting broadly consistent interaction experiences. User acceptance (UA), however, showed significant differences (F (2, 54) = 6.09, p = 0.004), with FGD2 demonstrating higher acceptance of technological solutions than FGD1.
Overall, these results indicate significant differences for mobile app usage, PU, PEU, and UA, suggesting that active NCAM farmers are more engaged and accepting of agricultural technologies. NCAM adoption, AI adoption willingness, and UX did not differ significantly, reflecting shared behaviours and perceptions across focus groups (Table 3).

6.2.2. Post Hoc Comparisons of Technology Adoption Variables

Post hoc Tukey HSD analyses were conducted to examine significant ANOVA results in detail (Table 4).
Perceived Usefulness (PU): Active NCAM farmers (FGD2) consistently scored higher than FGD1 across PU6, PU7, and PU8 (mean differences 0.79–0.85, p ≤ 0.002). FGD3 also scored higher than FGD1, though differences were smaller (0.37–0.43) and marginally significant. This suggests hands-on engagement enhances recognition of technological benefits, while interest without direct experience contributes to moderate perceived usefulness.
Perceived Ease of Use (PEU): Significant differences occurred only between FGD1 and FGD2 (PEU3–PEU5; mean differences 0.67–0.82, p ≤ 0.001), indicating that active practitioners find digital tools easier to use. FGD3 did not differ significantly from FGD1, suggesting awareness alone does not improve usability.
Mobile App Usage: FGD2 also reported higher app usage than FGD1 (mean difference = 1.57, p = 0.014), confirming that practical engagement correlates with digital adoption.
Overall, FGD2 consistently exhibited higher PU, PEU, and mobile app usage than FGD1. FGD3 showed elevated PU but no differences in PEU or app usage, highlighting a gap between awareness and practical adoption. UX did not vary across groups, indicating broadly positive experiences with communication channels regardless of prior exposure.
Notably, differences between FGD2 and FGD3 were not statistically significant, suggesting convergence in perceptions once farmers actively adopt NCAM practices. In contrast, user experience did not vary across groups, indicating that communication channels delivered broadly positive experiences regardless of prior exposure among farmers.

6.3. Discussion on Thematic Analysis

The discussions with farmers in the focus groups highlighted three key areas that shaped their thinking: (1) how they are practice Non-Chemical Agricultural Methods (NCAM) and how effective they consider them to be, (2) how they perceive the usefulness and scope of Artificial Intelligence (AI) in farming, and (3) what kind of information and communication strategies they rely on or expect. These themes show us how farmers talk about their own practices, how they respond to new technology, and what they need to make the adoption process smoother.

6.3.1. Discussion on Theme 1: Non-Chemical Agriculture Methods—Crops Cultivated in NCAM

When the farmers spoke about their crops, paddy the most frequently mentioned. A few referred to oilseeds and vegetables, and only one person mentioned fruits. In some groups, two farmers spoke about mixed cropping. Altogether, it was clear that NCAM farmers mainly focus on paddy, and while other crops grown lesser extent. As one farmer explained, “For us, paddy is the main crop. The government also treats it as the staple food crop of Tamil Nadu. They have given us Direct Purchase Centres (DPCs), so we can sell our paddy straight away without depending on middlemen.”
On effectiveness, farmers compared NCAM with chemical-based farming. One group strongly supported NCAM, saying it gave them healthier food and a sense of protecting the soil for the future. Another group pointed out that NCAM reduced their costs and helped biodiversity. A few were less engaged, but the overall mood was positive. Many stressed environmental benefits, especially soil health, while some expressed worries about how to sustain their families economically.
One farmer explained, “With this type of farming, our land is improving, soil is getting better, and biodiversity is returning. Even the cost of cultivation is less compared to chemical farming. But the problem is that the money in our hand is less. If we could sell directly, we might benefit immediately. When we market straight to the customer, profit is there but not in cash right away—that makes life hard for us. In DPC also, they treat our paddy just like any other paddy. They don’t give value for the variety or the breed we grow. Most of us are breed-focused farmers, but that effort is not recognized.”
The shortage of workforce was major concern. Across all groups, farmers agreed this was a serious problem. They acknowledged that NCAM needs more labour, but workers are not available. One participant noted, “People like to move forward to the city. Many of us do not have the reputation of being farmers, and many do not want the next generation to take this as their profession. They go into industry jobs. Now it is mostly the elder generation still in farming. Only if the next generation has their own interest—especially those with environmental concern—will they continue farming.” Farmers also said that the technologies they have are not enough to solve this labour gap, and they were not satisfied with the options available. At the same time, many described the labour-intensive nature of NCAM as something they value, since it keeps them actively involved and connected to the land.

6.3.2. Discussion on Theme 2: AI Technology Adoption

When AI was discussed, the farmers spoke in very practical terms. They were not very interested in high-tech tools but wanted to know how AI could save time, reduce workload, and protect crops. Weather alerts, crop monitoring, and automatic weeding were mentioned often. Irrigation advice also received attention. However, when asked about areas such as yield mapping or livestock monitoring, they showed little interest. “We are not very familiar with AI technology. Currently, we do not have access to tools like weather alerts, crop monitoring, or automatic weeding and irrigation. Perhaps in the future, we may adopt AI, especially due to workforce shortages, but this will depend on the feasibility of the technology.” one farmer explained.
Farmers expressed a strong preference for AI tools that directly reduce labour. Many said automation would help those most. A few were also interested in water-saving and climate-smart solutions, but overall, the emphasis was on practical, labour-saving applications.
When it came to learning and support, the farmers clearly said they prefer hands-on methods. Workshops and direct guidance from experts were mentioned most often. Digital platforms and apps were mentioned only a few times, and many said they don’t trust them as much. One farmer summed it up: “Instead of simply providing us with new technology, it is important that we first understand it. Many unnecessary technologies have been introduced, while many essential ones have been neglected. Technology becomes useful only when we are shown how to use it, when we experience it hands-on, and when it addresses our real needs. Only then can we accept and adopt it effectively.”

6.3.3. Discussion on Theme 3: Information Needs and Communication Strategy

Farmers were then asked how they get information. Most agreed that agricultural extension programmes are their main and most trusted source. They also mentioned farmer groups and community exchanges. Social media, mobile apps, and news were mentioned but less frequently.
The discussion around mobile apps, especially the government’s Uzhavar app, brought mixed reactions. Women farmers and smallholders in one group said they could not use the app because of the digital divide. One women farmer noted that, “Most of us do not have smartphones, and even if we do have we do not know how to use or follow everything on them. Traditionally, many farming tasks have been carried out by male farmers, who also take the key decisions in farming activities. If AI-based technologies such as robots are introduced, we may lose routine jobs like weeding, which could affect both our livelihood and the way farming roles are structured in our communities.” Others in mixed groups admitted they knew about the app but were dissatisfied with the limited or unclear messages. Middle age and more literate farmers were more positive, saying the app was helpful, but still required improvements. One participant remarked “The app is okay, we mainly use the Uzhavar App for communication and information from the Farmer Welfare Department. Nowadays, most of us are able to use smartphones, so if the app is improved with proper details, it will be more helpful.”
In terms of future communication, most farmers strongly felt that extension programmes should remain central. They trusted the face-to-face visits from officials and demonstrations in farmer schools. They said other channels like mobile apps, social media, and news have a role, but they cannot replace direct, personal engagement. As one farmer concluded, “We trust what we see in the field. An expert must show us how to use any new technology. Just watching on TV or mobile is not enough. When someone guides us and explains clearly, we can believe and take up the technology. It should also be affordable for us.”

6.4. Fuzzy-Set Qualitative Comparative Analysis (fsQCA)

The results of the fsQCA are outlined below, identifying the key conditions and combinations that are necessary and sufficient to drive AI adoption in agricultural practices.

6.4.1. Calibration of Raw Data

We calibrated raw survey data into fuzzy-set membership scores using direct calibration. This process assigned each case a score between 0.0 (Full Non-membership) and 1.0 (Full membership) for each condition. The three anchor points were:
  • Full Membership (1.0): High likelihood of AI adoption (e.g., strongly agree, high usage).
  • Crossover Point (0.5): Maximum ambiguity, neither in nor out.
  • Full Non-Membership (0.0): Low likelihood of AI adoption (e.g., strongly disagree, no usage).
For example, for ‘PEU16 (Use of agricultural mobile apps)’, a usage frequency of 10 was calibrated to 1.0, 5 to 0.5, and 0 to 0.0. The calibration criteria for all conditions are shown in Table 5.

6.4.2. Necessity Analysis

Necessity analysis identifies conditions that must be present for AI adoption (UA). The results (Table 5) show that three conditions have a perfect consistency score of 1.000: PEU14 (perception of NCAM as labour-intensive), PEU12 (workforce shortage), and PEU16 (use of agricultural mobile apps). This confirms they are necessary conditions, as all farmers who adopted AI shared these characteristics. Other conditions like BI19 (effectiveness of AI communication, 0.987) and PU9a (NCAM cost efficiency, 0.965) also showed high consistency but were not perfectly necessary can be seen in Table 6.

6.4.3. Sufficiency Analysis and Truth Table

The analysis then evaluated combinations of conditions that are sufficient to lead to AI adoption. A truth table was constructed listing all logical combinations of conditions. We set a frequency threshold of 2 (minimum number of cases per configuration) and a consistency threshold of 0.85 to identify sufficient paths. The truth table (Table 7) shows that adoption occurs consistently (outcome = 1) under specific configurations, most notably when PEU14, PEU12, and PEU16 are present together.
The truth table demonstrates different configurations leading to AI adoption.

6.4.4. Solution Pathways (fsQCA Results)

The fsQCA results produced three solutions. The parsimonious solution (consistency = 0.91) provides the core, essential conditions. The intermediate solution (consistency = 0.89) offers the best balance between detail and simplicity, revealing two primary pathways for AI adoption:
Pathway 1: Technologically Ready Farmers (PEU14, PEU12, PEU16, BI19 and PU9a).
Farmers adopt AI when they face labour pressures (PEU14, PEU12), are already using mobile apps (PEU16), find AI communication effective (BI19), and see NCAM as cost-efficient (PU9a).
Pathway 2: Institutionally Supported Farmers (PEU14, PEU12, PEU16~BI19, AU21a).
Farmers also adopt AI when they face the same labour pressures and use apps, but even if they find AI communication ineffective (~BI19), as long as they rely on agricultural extension programmes (AU21a).
The core conditions present in both pathways are PEU14, PEU12, and PEU16 confirming that labour pressure combined with digital readiness is a fundamental driver for AI adoption, as shown in Table 8.

6.5. Justice-Centred Discourse Analysis: Narratives on NCAM Adoption

6.5.1. Redistribution: Economic and Practical Considerations

Farmers’ narratives reveal how economic and material inequalities shape their engagement with Non-Chemical Agricultural Methods (NCAM). Discussions often highlight that NCAM is labour-intensive, shortage of workforce, and difficult to implement in the short term, even though many recognise its long-term benefits for sustainability and biodiversity. While farmers see NCAM as environmentally beneficial, it can be economically burdensome without support from institutions or appropriate technology. Farmers are generally open to new tools, such as mechanised irrigation and pest control, but remain uncertain about how these can be integrated into existing farming practices. These findings reflect issues of redistribution, where access to resources and costs influence the fairness and feasibility of adopting NCAM [58].

6.5.2. Recognition: Cultural Perceptions, Trust, and Social Influence

The analysis also emphasises the importance of recognition, where social identities and lived experience shape farmers’ willingness to adopt new methods. Discussions are often dominated by middle-aged male farmers, whose views carry authority due to their experience, while women and less prominently represented. Farming decisions are made by the male farmers and women farmers are considered as farm labourers rather than farmers. Women farmers are not considered as decision-makers in farm practice this reflects a non-recognition of their farming efforts, which needs to be addressed. When recognised, their knowledge and experience would foster confidence and willingness to engage in sustainable efforts and technology adoption. Comparatively, women farmers are less technology-driven, whereas male farmers are more information- and technology-driven with regard to NCAM practices. Many peer-group women farmers reported that they do not use smartphones. Older farmers tend to resist change, relying on traditional practices, whereas younger farmers show more openness. Trust is a key factor: farmers prefer learning through workshops, peer discussions, and government extension programmes. Their accounts highlight that recognition of their knowledge and experience is essential for adopting NCAM practices confidently [59].

6.5.3. Representation: Accessibility, Participation, and Behavioural Intention

Concerns about representation relate to how farmers can access and use new farming methods. While NCAM is described as labour-intensive, some farmers see potential benefits in using tools to reduce manual effort due to the increasing lack of workforce in the farming activities. However, barriers such as cost, access, complexity, and lack of training limit adoption. Farmers express doubts about whether they could manage new systems effectively, showing the need for inclusive training and support. Despite interest in innovations, farmers still rely on traditional extension services and face-to-face interactions rather than digital platforms. This suggests that farmers are not fully involved in decisions about how new practices are introduced or implemented; rather, they need information and communication support. Integrating education and participatory initiatives into existing extension programmes could help ensure that farmers’ voices are included in adopting NCAM [60].

6.5.4. Demographic Influences on Discourse (Cross-Cutting Dimension)

Demographic factors affect redistribution, recognition, and representation. Middle-aged male farmers dominate discussions, grounding conversations in tradition, while younger farmers are more receptive to innovations. These differences show how age, gender, and location shape farmers’ opportunities, whose perspectives are heard, and how resources and support are distributed fairly in promoting sustainable agriculture [61].

7. Key Findings

7.1. Thematic AnalysisFindings

Thematic analysis of the focus group discussions (FGDs) provides an in-depth understanding of farmers’ experiences, perceptions, and challenges regarding Non-Chemical Agricultural Methods (NCAM) and the adoption of Artificial Intelligence (AI) technologies. The analysis identifies four major themes—demographics, NCAM practices, AI technology adoption, and information and communication needs—situated within broader research on AI and sustainable agriculture in India and globally. The findings are interpreted using Nancy Fraser’s justice framework, encompassing economic, social, and representational dimensions of equity.

7.1.1. Demographics

The demographic profile reveals a skewed pattern in age and gender representation in agriculture. The absence of younger participants (below 30 years of age) in the focus group discussions (FGDs) indicates a generational gap that raises concerns about the continuity of sustainable farming practices. Furthermore, women were notably underrepresented—only 15 of 57 participants, primarily in FGD 1—highlighting a broader issue of social recognition. Despite their critical roles in agricultural labour and decision-making, women remain marginalised in predominantly male-dominated farming structures.

7.1.2. Theme 1: Non-Chemical Agricultural Methods (NCAM)

Farmers generally support non-chemical agricultural methods (NCAM), though significant challenges persist, notably workforce shortages and limited access to appropriate technologies. Priorities varied by group: FGD 2 emphasised cost efficiency, FGD 3 prioritised environmental sustainability, while FGD 1 demonstrated less engagement with the specific benefits of NCAM. Paddy cultivation dominates across all groups, with limited crop diversification due to economic constraints. Interestingly, the labour-intensive nature of NCAM was perceived positively—either because it creates employment opportunities or because it remains manageable under current conditions.
From an economic justice perspective, although farmers endorse NCAM, they often lack the institutional support and resources required for its full-scale implementation. Small-scale farmers, in particular, bear disproportionate labour burdens without sufficient technological or financial assistance.
AI and NCAM in India
Artificial Intelligence (AI) is increasingly recognised as a transformative force in agriculture, offering solutions that align with sustainable practices. When combined with Non-Chemical Agricultural Methods (NCAM)—including organic practices, precision irrigation, and targeted pest management—AI has the potential to enhance productivity while reducing chemical inputs and environmental impact [62]. Convolutional Neural Networks (CNNs) have become central to modern computer vision applications, enabling accurate pest detection, weed identification, and crop monitoring. However, in the Indian context, their effectiveness largely depends on the availability of high-quality, locally relevant datasets that reflect diverse crops, soils, and climatic conditions [63].
Globally, research highlights the significance of machine learning (ML), deep learning (DL), and time series analysis in strengthening agricultural decision-making. ML supports soil classification and crop selection, DL improves forecasting of yields and commodity prices, and time series approaches provide insights into demand and production trends, demonstrating their potential for improving food security and sustainability [64]. Recent studies in India extend these findings by showing how ML, DL, and hybrid models can be applied to crop selection, yield prediction, soil fertility classification, water management, and commodity price forecasting. Hybrid models integrating ML and DL have proven particularly effective for India’s smallholder-dominated farming systems, where localised solutions are essential. Furthermore, the development of multilingual translation tools and user-friendly digital platforms has been emphasised as critical for ensuring equitable access to AI-driven decision support across India’s diverse farming communities [65].
The emerging discourse on Agriculture 5.0 positions AI as a driver of automation, integrated farm systems, and improved supply chain linkages. While such technologies promise higher productivity and profitability, their adoption depends on building farmers’ awareness and strengthening institutional support. Scholars argue that transitioning toward Agriculture 5.0 requires a holistic approach that integrates advanced digital technologies with agro-ecology, agro-forestry, and livestock diversification, alongside sustained research and multi-stakeholder collaboration [66].

7.1.3. Theme 2: AI Technology Adoption

Farmers expressed interest in AI applications—particularly for weather alerts and automated weeding—yet actual usage and awareness remain limited. There is a marked preference for hands-on training and expert guidance over financial incentives, suggesting that informational barriers outweigh economic ones. The use of mobile-based tools and advanced technologies is especially limited among women and smallholder farmers.
This reflects both digital and representational inequalities. Despite clear interest, access to AI is hindered by insufficient digital infrastructure and farmers’ exclusion from the technology design process. The lack of participatory design undermines trust and inhibits adoption.
Successful Practices in India and Beyond
In India, AI is increasingly applied to weather forecasting, soil and groundwater monitoring, crop cycle management, and disease detection. These applications show promise in yield prediction, crop protection, and supply-chain efficiency, but adoption remains constrained by limited awareness and technical capacity in rural areas. Strengthening farmer training and institutional collaborations is key to scaling these practices sustainably [67].
Globally, AI supports semi-autonomous farming systems that enhance productivity while addressing climate change and food security challenges. Unlike past industrial and green revolutions that strained ecosystems, AI-driven frameworks—when paired with sustainable methods—offer pathways for resilient, eco-friendly agriculture. The COVID-19 pandemic highlighted the urgency of such technologies in building robust food systems [68].

7.1.4. Theme 3: Information Needs and Communication

Agricultural extension programmes remain the most trusted and frequently used sources of information across all FGDs. Reactions to mobile applications, such as the Uzhavar app, were mixed. Smallholder and women farmers in FGD 1 cited digital barriers, while participants in FGD 2 provided more positive feedback, though concerns about content sufficiency persisted. Across the board, farmers continue to favour face-to-face extension methods over digital platforms for future training and support.
This communication gap underscores challenges of both recognition and representation. Digital tools are not yet fully accessible or tailored to the diverse needs of farming communities, and farmers are rarely consulted during development. This lack of engagement exacerbates informational inequalities and deepens the digital divide.
Viewed through Nancy Fraser’s justice framework, the findings suggest that while farmers are willing to adopt both NCAM and AI technologies, their capacity to do so is constrained by systemic inequalities across dimension of access, recognition, and representation.
  • Economic injustice manifests through inadequate access to labour-saving technologies and financial support.
  • Social misrecognition affects women and smallholder farmers, whose knowledge and labour remain undervalued.
  • Representational exclusion persists, as farmers are seldom included in policy development or technology design.
To enable inclusive and sustainable farming, future initiatives must prioritise participatory training, culturally and linguistically appropriate communication, and farmer-led co-design of technologies. Only through these measures can NCAM and AI solutions promote not just productivity, but also equity, justice, and empowerment [69].
While these findings reflect the situation in India, similar challenges regarding digital access, literacy, and empowerment are observed globally, particularly among smallholder farmers in developing countries.
Smallholder farmers are crucial for rural livelihoods but face challenges such as climate variability, low productivity, high costs, limited credit, and constrained resources. AI and digital agriculture can address these challenges by providing predictive analytics, decision-support tools, and timely information. However, adoption remains limited globally, particularly in sub-Saharan Africa and parts of Asia, due to underdeveloped digital ecosystems, low digital literacy, affordability issues, and limited trust in data [70].
Successful digital transformation depends on inclusive service systems and organisational innovation rather than farm size. Building digital literacy and technical skills is essential for smallholders to effectively use digital platforms. Barriers such as data privacy concerns, labour replacement, and socio-political considerations further constrain adoption, highlighting the need for context-specific solutions [71].
Socio-cultural and power dynamics also influence adoption. While digital platforms improve communication with partners, smallholders often lack empowerment to negotiate support or flexibility, particularly under risk conditions like pest or disease outbreaks. Addressing these inequalities through equitable partnerships and restructured value chain interactions is critical to foster trust and inclusive use of AI [72].

7.2. Integrated Analysis of fsQCA and Discourse Findings

Sufficiency analysis indicated that several conditions exhibited high consistency (0.92–1.00), confirming their relevance for AI adoption, while coverage scores were moderate (0.4–0.9), suggesting that these conditions do not explain adoption across all cases (see Figure 4). Discourse analysis contextualised these findings by highlighting socio-cultural and structural factors that quantitative metrics alone could not capture.

7.2.1. Workforce Shortages

Workforce shortages were a highly consistent predictor of AI adoption (consistency = 0.95, coverage = 0.58). ANOVA revealed significant differences in perceived usefulness (PU) across groups (F (2, 105) = 4.87, p = 0.011), indicating that women smallholders reported lower PU due to disruption of informal labour-sharing systems. Discourse analysis confirmed that AI solutions were supportive by larger or better-resourced farmers but disruptive by smallholders reliant on traditional labour arrangements.

7.2.2. Institutional Trust and Extension Programmes

Extension programmes were identified as strong drivers of adoption (consistency = 0.93, coverage = 0.52). While PU differed significantly (F (2, 105) = 4.87, p = 0.011), perceptions of accessibility and engagement varied across groups. Discourse findings indicated that marginalised farmers received infrequent, generalised support, whereas better-resourced farmers benefitted from sustained, context-specific assistance, highlighting the mediating role of trust.

7.2.3. Mobile-Based Solutions

Mobile-based AI solutions demonstrated high consistency (0.92+) but moderate coverage (0.4–0.6). ANOVA results for mobile app usage showed significant group differences (F (2, 105) = 8.52, p = 0.014), with digitally literate farmers reporting higher ease of use. Discourse analysis attributed lower adoption among older or less digitally literate farmers to exclusion and limited training. The parsimonious fsQCA solution emphasised low-tech, community-based learning, confirming that adoption is most effective when aligned with existing knowledge systems rather than imposed top-down shown in Figure 5a,b and Figure 6.

7.2.4. Cost Efficiency and Attitudes Toward Use

Cost efficiency shaped attitudes toward use (A) (consistency = 0.92; coverage = 0.61). ANOVA revealed significant differences in user acceptance (UA) across groups (F (2, 105) = 6.09, p = 0.004), with marginalised farmers expressing lower optimism due to structural inequities. Farmers’ trust in AI depended on whether solutions complemented NCAM’s ecological and cultural practices.

7.2.5. Behavioural Intention and Social Validation

Behavioural intention (BI) was strongly influenced by social validation, with 60% of farmers citing peer influence as critical. Farmer-led networks achieved 70% coverage (consistency = 0.75), surpassing formal extension channels, highlighting the importance of horizontal, peer-driven knowledge exchange (Figure 7).

7.2.6. Perceived Ease of Use and User Experience

Perceived ease of use (PEU) differed significantly across groups (F (2, 105) = 3.94, p = 0.024), emphasising that training and digital literacy influence adoption. User experience (UX) differences were not statistically significant (F (2, 105) = 2.18, p = 0.119), suggesting that while interface familiarity matters, broader structural and cultural factors may be more influential in shaping adoption behaviours.

7.2.7. Linking TAM Constructs with Justice Dimensions

Mapping TAM variables onto Fraser’s justice framework revealed that:
  • Economic justice (redistribution) influenced PU and the capacity to act, as reflected in workforce shortages and cost barriers.
  • Cultural justice (recognition) affected PEU and attitudes, emphasising digital literacy, local knowledge, and alignment with traditional NCAM practices.
  • Political justice (representation) shaped BI, demonstrating the importance of inclusive decision-making, peer networks, and farmer-led initiatives.
These mappings are illustrated in Table 9, which operationalises TAM variables through Fraser’s justice dimensions in NCAM–AI adoption.
Collectively, these findings indicate that AI adoption in NCAM depends not only on technical performance but also on equity and contextual relevance. Integrating fsQCA, ANOVA, and discourse findings demonstrates that bridging technical robustness with inclusive, participatory approaches is essential for sustainable and equitable adoption.

8. Conclusions

This study applied the Integrated Mechanism for Sustainable Practices (IMSP)—framework to access systemic inequalities in the adoption of Non-Chemical Agricultural Methods (NCAM) and Artificial Intelligence (AI) in Tamil Nadu. It shows that the success of agricultural innovations depends not only on technological efficiency but equally on social inclusion, equity, and trust. The study specifically examined farmers’ perspectives on NCAM, AI, and access to agricultural information, addressing the research objectives. It highlights both the potential and limitations of NCAM and AI adoption, among marginalised farmers.
Farmers expressed strong support for NCAM, emphasising its positive impact on food quality, biodiversity, and soil health. However, uneven adoption persists, shaped by labour shortage, gendered roles, and market inequities.
This indicates that structural barriers, rather than farmers’ willingness, influence adoption. AI adoption is constrained by a persistent digital divide. Marginalised farmers reported lower awareness and acceptance, preferring simpler, locally adapted tools such as weather alerts. This reveals a justice gap between innovation and equitable access.
Access to agricultural information and trust in extension services also vary. Marginalised farmers received less tailored guidance, and many face difficulties using digital tools such as the Uzhavar app. This underlines the need for equitable communication strategies.
To ensure an equitable transition in agricultural technology, this study recommends:
  • Subsidise sustainable inputs and collective labour solution for NCAM.
  • Introduce AI literacy modules and establishing village-level technology hubs for real-time AI adoption support.
  • Ensure participatory governance where marginalised farmers have voice.
The justice-centred framework developed by Fraser’s principles of redistribution, recognition and representation with technology adoption constructs (usefulness, ease of use, attitude and behaviour intension).
The finding align with the Tamil Nadu Organic Farming Policy (2023) [73], and contribution to global sustainable debates. They resonate with SDG 2 (Zero Hunger), SDG 12 (Responsible Consumption and Production), and SDG 13 (Climate Action). Showing how locally grounded, equity-centred innovation pathways can have international relevance. This study focused on farmers’ perspective. Future research should include policymakers, technologist, and agricultural scientist, and explore how farm and regional variations affect adoption.
Overall, sustainable agricultural transformation requires robust, inclusive and justice-centred strategies. By integrating IMSP and TAM, this study provides a framework for ensuring that NCAM and AI benefit all farmers, especially the marginalised.

Author Contributions

Writing—original draft preparation, A.A.A.; supervision, I.A.A. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Indian Council of Social Science Research, Centrally Administered Full-Term Doctoral Fellowship/File No. RFD/2021-2022/GEN/MED/317.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board of Anna University, Chennai, India (Protocol Code: AU/Ethics/2021/093; Date of Approval: 15 March 2021).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study. Written informed consent was also obtained from the participants for publication of any potentially identifiable data.

Data Availability Statement

The data presented in this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.15874297.

Acknowledgments

The authors gratefully acknowledges the support of ICSSR for awarding the fellowship, which enabled the successful completion of this research. Sincere thanks are also extended to Nemili Block, and the farmers for their valuable guidance, cooperation, and participation.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Abbreviations

The following abbreviations are used in this manuscript:
AIArtificial Intelligence (AI)
APAIMSAndhra Pradesh Agriculture Information and Management System
BIBehavioural Intension
CNNConvolutional Neural Networks
CSAClimate-Smart Agriculture
DADiscourse Analysis
DPCDirect Purchase Centres
EVExternal Variable
FGDFocus Group Discussion
fsQCAFuzzy-set Qualitative Comparative Analysis
ICTInformation and Communication Technology
NCAMNon-Chemical Agricultural Methods
NMNFNational Mission on Natural Farming
PEUPerceived Ease of Use
PUPerceived Usefulness
SDGSustainable Development Goals
TAMTechnology Adoption Model

References

  1. Patel, S.K.; Sharma, A.; Singh, G.S. Traditional agricultural practices in India: An approach for environmental sustainability and food security. Energy Ecol. Environ. 2020, 5, 253–271. [Google Scholar] [CrossRef]
  2. Amalan, A.A.; Aram, I.A. Media influence in agriculture practices and scopes for non-chemical agricultural messages. Commun. Humanit. Soc. Sci. 2023, 3, 43–47. [Google Scholar] [CrossRef]
  3. Büyüközkan, G.; Uztürk, D. Integrated design framework for smart agriculture: Bridging the gap between digitalization and sustainability. J. Clean. Prod. 2024, 449, 141572. [Google Scholar] [CrossRef]
  4. Klerkx, L.; Jakku, E.; Labarthe, P. A review of social science on digital agriculture, smart farming and agriculture 4.0: New contributions and a future research agenda. NJAS Wagening. J. Life Sci. 2019, 90–91, 100315. [Google Scholar] [CrossRef]
  5. Finger, R. Digital innovations for sustainable and resilient agricultural systems. Eur. Rev. Agric. Econ. 2023, 50, 1277–1309. [Google Scholar] [CrossRef]
  6. The Times of India. Natural Farming Mission Encourages Ryots with Incentives, Cluster Support in Hassan Dist. The Times of India. 2025. Available online: https://timesofindia.indiatimes.com/city/mysuru/natural-farming-mission-encourages-ryots-with-incentives-cluster-support-in-hassan-dist/articleshow/122031053.cms?utm_source=chatgpt.com (accessed on 25 August 2025).
  7. The Times of India. Andhra Pradesh to Fully Digitize Agriculture from Kharif 2025. The Times of India. 2025. Available online: https://timesofindia.indiatimes.com/city/vijayawada/andhra-pradesh-to-fully-digitize-agriculture-from-kharif-2025/articleshow/121891066.cms?utm_source=chatgpt.com (accessed on 25 August 2025).
  8. Chaurasiya, P.; Kumar, S.; Singh, M.; Patel, V.K.; Singh, A.P.; Pandey, O.C.; Tewari, S. From tradition to technology: The impact of ai and IoT on agronomy practices in India and internationally. Int. J. Res. Agron. 2025, 8, 184–193. [Google Scholar] [CrossRef]
  9. Gyamfi, E.K.; Elsayed, Z.; Kropczynski, J.; Yakubu, M.A.; Elsayed, N. Agricultural 4.0 Leveraging on Technological Solutions: Study for Smart Farming Sector. In Proceedings of the 2024 IEEE 3rd International Conference on Computing and Machine Intelligence (ICMI), Mt Pleasant, MI, USA, 13–14 April 2024. [Google Scholar] [CrossRef]
  10. Bronson, K.; Knezevic, I. Big Data in food and agriculture. Big Data Soc. 2016, 3, 2053951716648174. [Google Scholar] [CrossRef]
  11. Howard, S.A. An Agricultural Testament; Oxford University Press: London, UK, 1943. [Google Scholar] [CrossRef]
  12. Liu, X.; Xu, Y.; Engel, B.A.; Sun, S.; Zhao, X.; Wu, P.; Wang, Y. The impact of urbanization and aging on food security in developing countries: The view from Northwest China. J. Clean. Prod. 2021, 292, 126067. [Google Scholar] [CrossRef]
  13. Saiz-Rubio, V.; Rovira-Más, F. From smart farming towards agriculture 5.0: A review on crop data management. Agronomy 2020, 10, 207. [Google Scholar] [CrossRef]
  14. Laskar, M.H. Examining the emergence of digital society and the digital divide in India: A comparative evaluation between urban and rural areas. Front. Sociol. 2023, 8, 1145221. [Google Scholar] [CrossRef]
  15. Gardezi, M.; Joshi, B.; Rizzo, D.M.; Ryan, M.; Prutzer, E.; Brugler, S.; Dadkhah, A. Artificial intelligence in farming: Challenges and opportunities for building trust. Agron. J. 2023, 116, 1217–1228. [Google Scholar] [CrossRef]
  16. Akash, J.H.; Aram, I.A. A convergent parallel mixed method of study for assessing the role of communication in community participation towards sustainable tourism. Environ. Dev. Sustain. 2022, 24, 12672–12690. [Google Scholar] [CrossRef]
  17. Zerfass, A.; Huck, S. Innovation, Communication, and Leadership: New Developments in Strategic Communication. Int. J. Strateg. Commun. 2007, 1, 107–122. [Google Scholar] [CrossRef]
  18. Spanaki, K.; Sivarajah, U.; Fakhimi, M.; Despoudi, S.; Irani, Z. Disruptive technologies in agricultural operations: A systematic review of AI-driven AgriTech research. Ann. Oper. Res. 2021, 308, 491–524. [Google Scholar] [CrossRef]
  19. Ahmed, I.; Jeon, G.; Piccialli, F. From artificial intelligence to Explainable Artificial Intelligence in Industry 4.0: A Survey on What, How, and Where. IEEE Trans. Ind. Inform. 2022, 18, 5031–5042. [Google Scholar] [CrossRef]
  20. Holzinger, A.; Fister, I.; Fister, I.; Kaul, H.P.; Asseng, S. Human-Centered AI in Smart Farming: Toward Agriculture 5.0. IEEE Access 2024, 12, 62199–62214. [Google Scholar] [CrossRef]
  21. Elbasi, E.; Mostafa, N.; Alarnaout, Z.; Zreikat, A.I.; Cina, E.; Varghese, G.; Shdefat, A.; Topcu, A.E.; Abdelbaki, W.; Mathew, S.; et al. Artificial Intelligence Technology in the Agricultural Sector: A Systematic Literature Review. IEEE Access 2023, 11, 171–202. [Google Scholar] [CrossRef]
  22. Chen, C.J.; Huang, Y.Y.; Li, Y.S.; Chang, C.Y.; Huang, Y.M. An AIoT Based Smart Agricultural System for Pests Detection. IEEE Access 2020, 8, 180750–180761. [Google Scholar] [CrossRef]
  23. Sharma, A.; Jain, A.; Gupta, P.; Chowdary, V. Machine Learning Applications for Precision Agriculture: A Comprehensive Review. IEEE Access 2021, 9, 4843–4873. [Google Scholar] [CrossRef]
  24. Zhou, Z.; Zahid, U.; Majeed, Y.; Nisha; Mustafa, S.; Sajjad, M.M.; Butt, H.D.; Fu, L. Advancement in artificial intelligence for on-farm fruit sorting and transportation. Front. Plant Sci. 2023, 14, 1082860. [Google Scholar] [CrossRef]
  25. Wu, X.; Aravecchia, S.; Lottes, P.; Stachniss, C.; Pradalier, C. Robotic weed control using automated weed and crop classification. J. Field Robot. 2020, 37, 322–340. [Google Scholar] [CrossRef]
  26. Gandhi, P.B.; Parejiya, A. The Power of Ai In Addressing The Challenges Faced By Indian Farmers In The Agriculture Sector: An Analysis. Tuijin Jishu/J. Propuls. Technol. 2023, 44, 4753–4777. [Google Scholar] [CrossRef]
  27. Tzachor, A. Barriers to AI Adoption in Indian Agriculture. Int. J. Innov. Digit. Econ. 2021, 12, 30–44. [Google Scholar] [CrossRef]
  28. Kumar, R.; Kumar, S.; Yashavanth, B.S.; Venu, N.; Meena, P.C.; Dhandapani, A.; Kumar, A. Natural Farming Practices for Chemical-Free Agriculture: Implications for Crop Yield and Profitability. Agriculture 2023, 13, 647. [Google Scholar] [CrossRef]
  29. Paudel, S.; Sah, L.P.; Devkota, M.; Poudyal, V.; Prasad, P.V.V.; Reyes, M.R. Conservation agriculture and integrated pest management practices improve yield and income while reducing labour, pests, diseases and chemical pesticide use in smallholder vegetable farms in Nepal. Sustainability 2020, 12, 6418. [Google Scholar] [CrossRef]
  30. Benbrook, C.; Kegley, S.; Baker, B. Organic farming lessens reliance on pesticides and promotes public health by lowering dietary risks. Agronomy 2021, 11, 1266. [Google Scholar] [CrossRef]
  31. Gomiero, T.; Pimentel, D.; Paoletti, M.G. Environmental impact of different agricultural management practices: Conventional vs. Organic agriculture. Crit. Rev. Plant Sci. 2011, 30, 95–124. [Google Scholar] [CrossRef]
  32. Sachithra, V.; Subhashini, L.D.C.S. How artificial intelligence uses to achieve the agriculture sustainability: Systematic review. Artif. Intell. Agric. 2023, 8, 46–59. [Google Scholar] [CrossRef]
  33. Andujar, D. Back to the Future: What Is Trending on Precision Agriculture? Agronomy 2023, 13, 2069. [Google Scholar] [CrossRef]
  34. Islam, M.; Bijjahalli, S.; Fahey, T.; Gardi, A.; Sabatini, R.; Lamb, D.W. Destructive and non-destructive measurement approaches and the application of AI models in precision agriculture: A review. Precis. Agric. 2024, 25, 1127–1180. [Google Scholar] [CrossRef]
  35. Thanki, H.; Shah, S.; Oza, A.; Vizureanu, P.; Burduhos-Nergis, D.D. Sustainable Consumption: Will They Buy It Again? Factors Influencing the Intention to Repurchase Organic Food Grain. Foods 2022, 11, 3046. [Google Scholar] [CrossRef]
  36. Davis, F.D. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Q. 1989, 13, 319–340. [Google Scholar] [CrossRef]
  37. Pierpaoli, E.; Carli, G.; Pignatti, E.; Canavari, M. Drivers of Precision Agriculture Technologies Adoption: A Literature Review. Procedia Technol. 2013, 8, 61–69. [Google Scholar] [CrossRef]
  38. Venkatesh, V.; Davis, F.D. Theoretical extension of the Technology Acceptance Model: Four longitudinal field studies. Manag. Sci. 2000, 46, 186–204. [Google Scholar] [CrossRef]
  39. Marangunić, N.; Granić, A. Technology acceptance model: A literature review from 1986 to 2013. Univers. Access Inf. Soc. 2014, 14, 81–95. [Google Scholar] [CrossRef]
  40. Mohr, S.; Kühl, R. Acceptance of artificial intelligence in German agriculture: An application of the technology acceptance model and the theory of planned behavior. Precis. Agric. 2021, 22, 1816–1844. [Google Scholar] [CrossRef]
  41. Bagheri, A.; Bondori, A.; Allahyari, M.S.; Surujlal, J. Use of biologic inputs among cereal farmers: Application of technology acceptance model. Environ. Dev. Sustain. 2021, 23, 5165–5181. [Google Scholar] [CrossRef]
  42. Braun, V.; Clarke, V. Using thematic analysis in psychology. Qual. Res. Psychol. 2006, 3, 77–101. [Google Scholar] [CrossRef]
  43. Kress, G. Critical Discourse Analysis. Annu. Rev. Appl. Linguist. 1991, 11, 84–99. [Google Scholar] [CrossRef]
  44. Phillips, N.; Lawrence, T.B.; Hardy, C. Discourse and institutions. Acad. Manag. Rev. 2004, 29, 636–652. [Google Scholar] [CrossRef]
  45. Ragin, C.C. Redesigning Social Inquiry: Fuzzy Sets and Beyond; University of Chicago Press: Chicago, IL, USA, 2008. [Google Scholar]
  46. Mondal, S.; Das, S.; Vrana, V.G. Exploring the Role of Artificial Intelligence in Achieving a Net Zero Carbon Economy in Emerging Economies: A Combination of PLS-SEM and fsQCA Approaches to Digital Inclusion and Climate Resilience. Sustainability 2024, 16, 10299. [Google Scholar] [CrossRef]
  47. Mustafa, S.; Zhang, W.; Shehzad, M.U.; Anwar, A.; Rubakula, G. Does Health Consciousness Matter to Adopt New Technology? An Integrated Model of UTAUT2 with SEM-fsQCA Approach. Front. Psychol. 2022, 13, 836194. [Google Scholar] [CrossRef]
  48. Cubric, M. Drivers, barriers and social considerations for AI adoption in business and management: A tertiary study. Technol. Soc. 2020, 62, 101257. [Google Scholar] [CrossRef]
  49. Fraser, N. Debate: Recognition or Redistribution? A Critical Reading of Iris Young’s Justice and the Politics of Difference. J. Polit. Philos. 1995, 3, 166–180. [Google Scholar] [CrossRef]
  50. Nyumba, T.O.; Wilson, K.; Derrick, C.J.; Mukherjee, N. The use of focus group discussion methodology: Insights from two decades of application in conservation. Methods Ecol. Evol. 2018, 9, 20–32. [Google Scholar] [CrossRef]
  51. Ryan, G.W.; Bernard, H.R. Techniques to Identify Themes. Field Methods 2003, 15, 85–109. [Google Scholar] [CrossRef]
  52. Henson, R.N. Analysis of Variance (ANOVA). Brain Mapp. 2015, 1, 477–481. [Google Scholar] [CrossRef]
  53. Lochmiller, C.R. Conducting thematic analysis with qualitative data. Qual. Rep. 2021, 26, 2029–2044. [Google Scholar] [CrossRef]
  54. Wood, B.A.; Blair, H.T.; Gray, D.I.; Kemp, P.D.; Kenyon, P.R.; Morris, S.T.; Sewell, A.M. Agricultural science in the wild: A social network analysis of farmer knowledge exchange. PLoS ONE 2014, 9, e105203. [Google Scholar] [CrossRef]
  55. Chindasombatcharoen, N.; Tsolakis, N.; Kumar, M.; O’Sullivan, E. Navigating psychological barriers in agricultural innovation adoption: A multi-stakeholder perspective. J. Clean. Prod. 2024, 475, 143695. [Google Scholar] [CrossRef]
  56. McGuire, E.; Al-Zu’bi, M.; Boa-Alvarado, M.; Luu, T.T.G.; Sylvester, J.M.; Leñero, E.M.V. Equity principles: Using social theory for more effective social transformation in agricultural research for development. Agric. Syst. 2024, 218, 103999. [Google Scholar] [CrossRef]
  57. Kraus, S.; Ribeiro-Soriano, D.; Schüssler, M. Fuzzy-set qualitative comparative analysis (fsQCA) in entrepreneurship and innovation research—The rise of a method. Int. Entrep. Manag. J. 2018, 14, 15–33. [Google Scholar] [CrossRef]
  58. Fraser, N. From redistribution to recognition? Dilemmas of justice in a post-socialist age. In Feminism And Politics: Oxford Readings In Feminism; Oxford Academi: Oxford, UK, 1995. [Google Scholar] [CrossRef]
  59. Fraser, N. Interview with Nancy Fraser: Justice as Redistribution, Recognition, and Representation. MR Online. 2009. Available online: https://mronline.org/2009/05/16/interview-with-nancy-fraser-justice-as-redistribution-recognition-and-representation/ (accessed on 22 August 2025).
  60. Petoukhov, K. Recognition, Redistribution, and Representation: Assessing the Transformative Potential of Reparations for the Indian Residential Schools Experience. McGill Sociol. Rev. 2013, 3, 73–91. [Google Scholar]
  61. Fraser, N. Social Justice in the Age of Identity Politics: Redistribution, Recognition, and Participation; Wissenschaftszentrum Berlin für Sozialforschung: Berlin, Germany, 1998. [Google Scholar]
  62. Talaviya, T.; Shah, D.; Patel, N.; Yagnik, H.; Shah, M. Implementation of artificial intelligence in agriculture for optimisation of irrigation and application of pesticides and herbicides. Artif. Intell. Agric. 2020, 4, 58–73. [Google Scholar] [CrossRef]
  63. Dhanya, V.G.; Subeesh, A.; Kushwaha, N.L.; Vishwakarma, D.K.; Nagesh Kumar, T.; Ritika, G.; Singh, A.N. Deep learning based computer vision approaches for smart agricultural applications. Artif. Intell. Agric. 2022, 6, 211–229. [Google Scholar] [CrossRef]
  64. Ali, Z.; Muhammad, A.; Lee, N.; Waqar, M.; Lee, S.W. Artificial Intelligence for Sustainable Agriculture: A Comprehensive Review of AI-Driven Technologies in Crop Production. Sustainability 2025, 17, 2281. [Google Scholar] [CrossRef]
  65. Akkem, Y.; Biswas, S.K.; Varanasi, A. Smart farming using artificial intelligence: A review. Eng. Appl. Artif. Intell. 2023, 120, 105899. [Google Scholar] [CrossRef]
  66. Vinod, C.S.S.; Anand, H.S.; Albaaji, G.F. Precision farming for sustainability: An agricultural intelligence model. Comput. Electron. Agric. 2024, 226, 109386. [Google Scholar] [CrossRef]
  67. Naresh, R.K.; Chandra, M.S.; Vivek; Shivangi; Charankumar, G.R.; Chaitanya, J.; Alam, M.S.; Singh, P.K.; Ahlawat, P. The Prospect of Artificial Intelligence (AI) in Precision Agriculture for Farming Systems Productivity in Sub-Tropical India: A Review. Curr. J. Appl. Sci. Technol. 2020, 39, 96–110. [Google Scholar] [CrossRef]
  68. Bhagat, P.R.; Naz, F.; Magda, R. Artificial intelligence solutions enabling sustainable agriculture: A bibliometric analysis. PLoS ONE 2022, 17, e0268989. [Google Scholar] [CrossRef]
  69. Honneth, A.; Fraser, N. Redistribution or Recognition? A Political-Philosophical Exchange; Verso Books: London, UK, 2003; Volume 74, pp. 1337–1346. [Google Scholar]
  70. Gumbi, N.; Gumbi, L.; Twinomurinzi, H. Towards Sustainable Digital Agriculture for Smallholder Farmers: A Systematic Literature Review. Sustainability 2023, 15, 12530. [Google Scholar] [CrossRef]
  71. Xie, L.; Luo, B.; Zhong, W. How are smallholder farmers involved in digital agriculture in developing countries: A case study from China. Land 2021, 10, 245. [Google Scholar] [CrossRef]
  72. Agyekumhene, C.; De Vries, J.; van Paassen, A.; Schut, M.; MacNaghten, P. Making smallholder value chain partnerships inclusive: Exploring digital farm monitoring through farmer friendly smartphone platforms. Sustainability 2020, 12, 4580. [Google Scholar] [CrossRef]
  73. Agriculture-Farmers Welfare Department. Tamil Nadu Organic Farming Policy. 2023. Available online: https://agritech.tnau.ac.in/pdf/66617733-Tamil-Nadu-Organic-Farming-Policy-2023_230315_093042.pdf (accessed on 12 January 2025).
Figure 1. Technology Acceptance Model flow chart.
Figure 1. Technology Acceptance Model flow chart.
Sustainability 17 08865 g001
Figure 2. Conceptual Flow Chart: Integrated Mechanism for Sustainable Agricultural Practices.
Figure 2. Conceptual Flow Chart: Integrated Mechanism for Sustainable Agricultural Practices.
Sustainability 17 08865 g002
Figure 3. Focus group discussion with farmers from Nemili Block and Thuraiyur Village.
Figure 3. Focus group discussion with farmers from Nemili Block and Thuraiyur Village.
Sustainability 17 08865 g003
Figure 4. Sufficiency Analysis of AI Adoption—Consistency and Coverage Scores.
Figure 4. Sufficiency Analysis of AI Adoption—Consistency and Coverage Scores.
Sustainability 17 08865 g004
Figure 5. (a) Consistency scores ranging from 0.800 to 1.000, with the parsimonious solution achieving a perfect score, highlight the technical robustness of AI adoption drivers such as workforce shortage; (b) Coverage scores underscore that these high-consistency conditions do not apply universally.
Figure 5. (a) Consistency scores ranging from 0.800 to 1.000, with the parsimonious solution achieving a perfect score, highlight the technical robustness of AI adoption drivers such as workforce shortage; (b) Coverage scores underscore that these high-consistency conditions do not apply universally.
Sustainability 17 08865 g005
Figure 6. Reinforces that real-world uptake varies by demographic and structural context.
Figure 6. Reinforces that real-world uptake varies by demographic and structural context.
Sustainability 17 08865 g006
Figure 7. Illustration of the TAM Pathway.
Figure 7. Illustration of the TAM Pathway.
Sustainability 17 08865 g007
Table 1. The table represents FGD insights reorganised for fsQCA analysis, using TAM in the context of NCAM and AI adoption.
Table 1. The table represents FGD insights reorganised for fsQCA analysis, using TAM in the context of NCAM and AI adoption.
S. No.TAM CategoryQuestion/Indicator
1External VariableAge
2External VariableGender
3External VariableDistrict
4External VariableBlock
5External VariableVillage
6External VariableFarming Experience
7External VariableHave you adopted Non-Chemical Agricultural Methods (NCAM)?
8a–8fExternal VariableWhat crops do you cultivate using NCAM?
(a) Paddy
(b) Vegetables
(c) Fruits
(d) Oil Seeds
(e) Pulses
(f) All Crops
9a–9fPerceived Usefulness (PU)Effectiveness of NCAM vs. chemical farming:
(a) Cost Efficiency
(b) Good Food
(c) Soil & Water Quality
(d) Biodiversity
(e) Sustainable Economy
(f) All the Above
10a–10gPerceived Usefulness (PU)AI technologies perceived as useful in NCAM:
(a) Weather Alert
(b) Crop & Soil Monitoring
(c) Pest & Disease Detection
(d) Automated Weeding & Harvesting
(e) Automated Irrigation
(f) Yield Mapping
(g) Livestock Health Monitoring
11a–11ePerceived Usefulness (PU)Support needed to adopt AI in NCAM:
(a) Workshops
(b) Expert Advice
(c) AI Equipment
(d) Mobile Apps
(e) Govt Support
12Perceived Ease of Use (PEU)Do you have a shortage of workforce?
13Perceived Ease of Use (PEU)Do you have sufficient technology to balance the workforce?
14Perceived Ease of Use (PEU)Do you consider NCAM to be labour-intensive?
15a–15fPerceived Ease of Use (PEU)Challenges in adopting AI in farming:
(a) Automation & Robotics
(b) Irrigation Management
(c) Climate-Smart Agriculture
(d) Data & Risk Management
(e) Crop & Soil Health Analysis
(f) None of the Above
16Perceived Ease of Use (PEU)Do you use any agricultural mobile apps for farming guidance?
17Attitude Toward Using (A)Are you aware of AI applications in farming?
18Attitude Toward Using (A)Are you willing to adopt AI-driven technologies in your farm?
19Behavioural Intention (BI)How effective are communication channels in providing relevant AI information?
20a–20eBehavioural Intention (BI)Preferred channels for AI communication:
(a) Extension Programmes
(b) Farm Schools/Groups
(c) News/Media
(d) Mobile Apps
(e) Social media
21a–21eActual System UsePrimary sources for NCAM & farming knowledge:
(a) Agricultural Extension Programmes
(b) Farmer Groups
(c) Social media
(d) Agricultural Apps
(e) News/Media
22Actual System UseDo you use any agricultural mobile apps for farming guidance?
Table 2. Descriptive Statistics of all three Focus Group Participants (n = 57).
Table 2. Descriptive Statistics of all three Focus Group Participants (n = 57).
VariableMeanMedianStandard Deviation
Demographics & Farming Experience
Age4.2940.77
Gender1.2610.44
District21.17226.48
Occupation110.00
Farming Experience4.3550.99
NCAM Adoption & Labour Factors
Adoption to NCAM1.9211.00
Shortage of Workforce1.0510.29
Level of Sufficient Technology4.8450.64
Level of Labour Intensiveness in NCAM1.0010.00
AI Awareness & Adoption
Awareness on AI4.6251.00
Willingness to Adopt AI2.1731.18
Communication & ICT Usage
Effectiveness of Communication Channels1.0810.28
Mobile Application Usage3.6351.83
Table 3. ANOVA Results across Farmer Focus Groups.
Table 3. ANOVA Results across Farmer Focus Groups.
Dependent VariableSum of Squares (Between Groups)dfMean SquareF-Valuep-ValueInterpretation
NCAM Adoption2.1221.061.210.305Not Significant
AI Adoption Willingness1.9520.9750.910.412Not Significant
Mobile App Usage18.3729.1858.520.014Significant
Perceived Usefulness (PU)12.4126.214.870.011Significant
Perceived Ease of Use (PEU)10.3625.183.940.024Significant
User Experience (UX)8.9524.472.180.119Not Significant
User Acceptance (UA)14.7227.366.090.004Significant
Table 4. Post hoc comparisons (Tukey HSD) across farmer focus groups.
Table 4. Post hoc comparisons (Tukey HSD) across farmer focus groups.
VariableGroup ComparisonMean Differencep-ValueInterpretation
PU6FGD1 vs. FGD2−0.790.001FGD2 > FGD1
PU6FGD1 vs. FGD3−0.430.046FGD3 > FGD1
PU7FGD1 vs. FGD2−0.800.000FGD2 > FGD1
PU7FGD1 vs. FGD3−0.370.041FGD3 > FGD1
PU8FGD1 vs. FGD2−0.850.002FGD2 > FGD1
PU8FGD1 vs. FGD3−0.410.049FGD3 > FGD1
PEU3FGD1 vs. FGD2−0.670.001FGD2 > FGD1
PEU4FGD1 vs. FGD2−0.820.000FGD2 > FGD1
PEU5FGD1 vs. FGD2−0.710.001FGD2 > FGD1
Mobile App UsageFGD1 vs. FGD2−1.570.014FGD2 > FGD1
Table 5. Calibration of key conditions for AI adoption.
Table 5. Calibration of key conditions for AI adoption.
ConditionRaw Data RangeFull Membership (1.0)Crossover (0.5)Full Non-Membership (0.0)
PEU14 (Labour-intensive NCAM perception)1–5 (Likert Scale)531
PEU12 (Workforce shortage)1–5531
PEU16 (Use of agricultural mobile apps)0–10 (usage frequency)1050
BI19 (Effectiveness of AI communication)1–5531
PU9a (NCAM cost efficiency)1–5531
AU21a (Agricultural extension programme info source)1–5531
PU9d (NCAM biodiversity benefits)1–5531
BI20a (AI communication via extension programmes)1–5531
PU9b (NCAM impact on food quality)1–5531
EV2 (Gender) (Male = 1, Female = 0)0–11 (Male)0.5
Table 6. Conditions present in AI adoption cases.
Table 6. Conditions present in AI adoption cases.
ConditionConsistencyCoverage
PEU14 (Labour-intensive perception of NCAM)1.0000.339
PEU12 (Workforce shortage)1.0000.344
PEU16 (Use of agricultural mobile apps)1.0001.000
BI19 (Effectiveness of AI communication)0.9870.342
PU9a (NCAM cost efficiency)0.9650.405
AU21a (Agricultural extension programme as information source)0.9480.346
PU9d (NCAM biodiversity benefits)0.9480.407
BI20a (Extension programmes for AI communication)0.9480.352
PU9b (NCAM impact on food quality)0.9480.339
EV2 (Gender)0.9220.360
Table 7. Configurations leading to AI adoption.
Table 7. Configurations leading to AI adoption.
PEU14PEU12PEU16BI19PU9aAU21aCasesAI Adoption Rate
1110.751111
11111011
111111151
10.500.750010
10.670.511110
1100.750120
11010170
11011020
110111240
Table 8. Balance between complexity and simplification.
Table 8. Balance between complexity and simplification.
Solution TypeConsistencyCoverage
Complex0.870.78
Parsimonious0.910.82
Intermediate0.890.80
Table 9. Operationalising TAM Variables through Fraser’s Justice Dimensions in NCAM–AI Adoption.
Table 9. Operationalising TAM Variables through Fraser’s Justice Dimensions in NCAM–AI Adoption.
TAM VariableEconomic Justice (Redistribution)Cultural Justice (Recognition)Political Justice (Representation)
Perceived Usefulness (PU)High consistency of workforce shortages (0.95) shows AI’s labour-saving potential, but coverage gaps reveal inequities in access to affordable solutions.Traditional labour-sharing norms disrupted, making women smallholders perceive AI as less supportive.Limited inclusion of smallholders in extension programmes reduces alignment of AI tools with their needs.
Perceived Ease of Use (PEU)Mobile literacy (0.88) strongly influences adoption; cost barriers restrict training and access.Older and less digitally literate farmers excluded, creating cultural divides in usability.Training formats are top-down; lack of farmer participation in design limits accessibility.
Attitude Toward Use (A)Cost efficiency (0.92) shapes positive attitudes where resources are available, but structural inequities dampen optimism among marginalised groups.Farmers’ trust depends on whether AI complements NCAM’s ecological and cultural values.Generalised extension support fails to represent diverse farmer contexts, reducing credibility.
Behavioural Intention (BI)Market access barriers restrict farmers’ capacity to act on positive intentions.Peer influence cited by 60% of farmers shows social validation as critical to intention.Farmer-led networks (70% coverage) outperform formal institutions, highlighting representation through collective action.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Amalan, A.A.; Aram, I.A. Artificial Intelligence Adoption in Non-Chemical Agriculture: An Integrated Mechanism for Sustainable Practices. Sustainability 2025, 17, 8865. https://doi.org/10.3390/su17198865

AMA Style

Amalan AA, Aram IA. Artificial Intelligence Adoption in Non-Chemical Agriculture: An Integrated Mechanism for Sustainable Practices. Sustainability. 2025; 17(19):8865. https://doi.org/10.3390/su17198865

Chicago/Turabian Style

Amalan, Arokiaraj A., and I. Arul Aram. 2025. "Artificial Intelligence Adoption in Non-Chemical Agriculture: An Integrated Mechanism for Sustainable Practices" Sustainability 17, no. 19: 8865. https://doi.org/10.3390/su17198865

APA Style

Amalan, A. A., & Aram, I. A. (2025). Artificial Intelligence Adoption in Non-Chemical Agriculture: An Integrated Mechanism for Sustainable Practices. Sustainability, 17(19), 8865. https://doi.org/10.3390/su17198865

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop