1. Introduction
Over the years, sustainable agriculture has evolved into a global necessity, driven by the need to address the growing environmental impact of traditional farming practices. This impact, which affects not only the environment but also society and economies worldwide, demands urgent attention. In this context, artificial intelligence (AI) emerges as an innovative tool with vast potential to revolutionize traditional agricultural practices (Mana et al., 2024). AI can optimize existing resources and implement solutions that mitigate the negative effects of agriculture on the environment. Sustainable agriculture seeks a balance between society, the economy, and the environment, promoting the conservation of resources, reducing unnecessary use of materials, and minimizing environmental harm. As the demand for food increases, the agricultural sector must respond proportionately to external factors such as climate change and environmental degradation (Okot & Ojok, 2023).
The digital era has significantly impacted every aspect of modern life, and agriculture is no exception. In this scenario, sustainable agriculture benefits from the application of information technologies, offering an alternative to traditional farming and expanding the possibilities for improvement in the agricultural sector. Digital transformation in agriculture is fueled by the need to address challenges such as population growth, environmental impact, resource scarcity, and the rising demand for food sustainably and efficiently (Calicioglu et al., 2019). According to ElMassah & Mohieldin (2020), digital transformation is driven by the pressing need to tackle these challenges. In this context, AI offers innovative solutions through predictive and descriptive methods, as well as automated tools designed for specific agricultural tasks. AI can help farmers monitor and analyze variables involved in agricultural activities, leading to more efficient and sustainable practices.
Bertoglio et al. (2021) identified five key trends in the application of information technologies in sustainable agriculture: climate-smart agriculture, site-specific management, remote sensing, the Internet of Things (IoT), and AI. These technologies aim to predict and describe the behavior of variables critical to agricultural activities, optimizing both productivity and sustainability. However, there remains a gap in the literature that comprehensively addresses the impact of the digital age on sustainable agriculture. As countries transition toward more sustainable practices, it becomes increasingly important to understand how digital advancements affect the development of sustainable farming.
While digitalization has expanded access to information and empowered individuals and society, there is still a significant digital divide in the agricultural sector (Jamil, 2021). Factors such as geographic location and limited internet access, combined with the lack of academic preparation among farmers, represent major challenges to the adoption of digital technologies. According to Autor (2019), the delay in the implementation of digital technologies and the lack of a competent workforce in agriculture are due to skill gaps compared to other sectors. Geographic, social, and economic conditions often influence these disparities.
The role of AI in sustainable agriculture
In a global context where agriculture represents a vital sector for the population, the need for sustainable practices has grown more urgent. Rapid population growth, climate change, and resource scarcity pose significant challenges to traditional agricultural practices (Goel et al., 2021). Digital solutions, especially AI, have emerged as innovative and necessary tools to address these challenges.
Sustainable agriculture is an integrated system designed to meet the current and future needs of the population. According to DeLonge et al. (2016), it is essential to produce enough food while ensuring long-term benefits, correct resource management, and improved quality of life for both farmers and the general population. This system operates holistically rather than as isolated components. The growing focus on sustainability and digital transformation in agriculture, therefore, emphasizes the importance of AI as a catalyst for improving agricultural practices. Sustainable agriculture aims to maintain productivity over the long term while managing natural resources responsibly. Muhie (2022) emphasizes that sustainability in agriculture seeks to preserve productivity by optimizing resource use and reducing unnecessary consumption to safeguard the environment.
However, agriculture also contributes significantly to global challenges. According to Khatri et al. (2024), current food systems generate considerable threats to human health and the environment, causing elevated levels of pollution and waste. Alarmingly, one-third of global food production is wasted, exacerbating food insecurity and overshadowing efforts to meet climate goals. The World Bank also notes that agriculture contributes 30% of global greenhouse gas emissions, highlighting the urgency of adopting sustainable practices to mitigate environmental degradation.
The digital age has transformed agriculture, shifting it from traditional methods to modern, technology-driven approaches that enhance efficiency and sustainability. Technologies like AI and IoT are reshaping how agricultural tasks are managed, optimizing resource usage and improving overall productivity. According to Wessel et al. (2021), digital transformation involves the integration of information technology across all aspects of an organization to add value. Examples of this transformation include developing digital solutions like mobile applications and e-commerce platforms, migrating IT infrastructure to the cloud, and implementing smart sensors to reduce operational costs. Veeramanju (2023), further explains that digital transformation in agriculture incorporates AI to enhance information optimization, improve workflow automation, and support real-time decision-making. This transformation not only streamlines operations but also generates new business opportunities by facilitating more efficient resource management and improving food security.
AI has emerged as a key technology in sustainable agriculture, providing tools that can optimize resource use, predict environmental factors, and automate critical tasks. According to Galiana et al. (2024), AI is not merely an advanced technology but a driver of ethical and moral transformations in how societies address environmental challenges. Early developments in AI in the 1950s laid the groundwork for its current applications, where machine learning algorithms and predictive models are instrumental in enhancing agricultural efficiency. AI’s ability to learn from data and provide actionable insights makes it particularly valuable in agriculture. As Alassery et al. (2022) explain, AI’s specialized algorithms enable a successful alliance between energy design and current technology, providing tools for monitoring, detecting, and solving various issues related to agriculture. These algorithms help farmers make better decisions, such as optimizing water use, predicting crop yields, and monitoring soil health, thus reducing resource waste and improving sustainability.
AI’s integration into agriculture extends beyond productivity improvements; it also plays a crucial role in environmental conservation. According to Lukacz (2024), Microsoft’s AI initiatives aim to bridge gaps in environmental monitoring and management, enhancing conservation efforts and mitigating climate impact. AI applications can forecast water consumption, predict climate variability, and offer solutions that promote sustainable farming practices.
Shen et al. (2024) highlights the importance of multi-tasking AI models in improving agricultural sustainability. For example, AI can interpret diverse data types, such as weather patterns and satellite imagery, to predict environmental phenomena. These capabilities help farmers make informed decisions, particularly in the face of unpredictable climate events like droughts or floods. AI also enhances the accuracy of environmental monitoring using sensors and satellite imagery. Sensors provide continuous data on temperature, wind, and other environmental factors, while satellite images offer broader perspectives on environmental changes. By combining both types of data, AI provides a complete and more precise picture of environmental processes, allowing for better decision-making in managing climate risks.
The ongoing Industry 4.0 is reshaping human interactions with the environment, particularly in agriculture. This revolution is characterized by the integration of AI and machine learning, which processes vast amounts of information in real time to develop innovative solutions. According to Ashima et al. (2021), Industry 4.0 promotes greater efficiency in agriculture through automation and the use of smart devices, helping manufacturers meet consumer demands through mass customization while improving decision-making processes.
Pereira & Romero (2017) describe Industry 4.0 as a technological revolution that changes how organizations operate, design, produce, and supply goods and services. In agriculture, this revolution integrates advanced technologies like AI, IoT, and Big Data, creating systems that communicate and work together to improve overall effectiveness. By adopting these technologies, agricultural enterprises can better meet the growing demand for food while addressing the need for sustainable resource management.
Overcoming barriers to digital transformation in agriculture
While digital transformation offers many benefits, it also risks deepening existing inequalities. Helsper (2021) describes how digitalization has generated new opportunities but also widened the gap between those who fully adopt innovative technologies and those who resist or are unaware of them. Chetty et al. (2018) explore the digital divide in the context of Industry 4.0, identifying three key dimensions: access, practical use, and enabling use. Access refers to the availability of devices and internet connectivity, often limited by socioeconomic, gender, and territorial factors. In an era where technology is critical to agricultural success, the lack of access to digital tools becomes a significant barrier to participation in an increasingly digitalized economy.
The divide in practical use refers to differences in how people utilize digital devices and the internet, while the enabling use dimension focuses on meaningful engagement with technology, such as creating and sharing data. Overcoming these gaps is crucial to ensuring that all farmers, regardless of their location or socioeconomic status, can benefit from AI and other digital technologies in sustainable agriculture.
The adoption of AI and other digital technologies in agriculture faces several barriers. One of the main challenges is the lack of infrastructure and access to technology in rural areas. According to Rayna & Striukova (2021), digital technologies are essential for promoting innovation, growth, and development. However, these technologies can also amplify existing social, economic, and territorial inequalities.
Addressing these challenges requires a concerted effort to develop infrastructure, provide access to education, and create policies that promote the equitable distribution of digital tools. Without these efforts, the potential of AI to transform agriculture will be limited by the inability of many farmers to fully utilize these technologies. The integration of AI in sustainable agriculture presents a transformative opportunity to address global challenges such as climate change, food security, and resource scarcity. However, this transformation is not without its obstacles. The digital divide, lack of infrastructure, and social inequalities must be addressed to ensure that all realize AI’s benefits. With strategic investments and policies, AI can revolutionize agriculture, making it more sustainable, efficient, and resilient in the face of future challenges.
In Costa Rica, agriculture is a key economic driver, contributing significantly to the nation's economy. The Ministry of Agriculture and Livestock’s 2023-2024 report highlights that the agricultural sector contributed over 5 billion USD, accounting for 67% of Costa Rica's gross agricultural production (Parada Gómez & Jiménez Ureña, 2023). AI offers an opportunity to address the blind spots in Costa Rican agriculture, where specific execution times are required to maintain the quality of agricultural products. Innovation in this area should align with national sustainability commitments, ensuring that productivity is maintained without compromising the environment.
The transition toward sustainable agriculture presents an opportunity to address current environmental and socioeconomic challenges. Climate-smart practices enhance the adaptability and sustainability of the agricultural sector, allowing it to better respond to these challenges (Azadi et al., 2021). However, the pursuit of sustainability through information technologies involves significant economic investment, particularly in the implementation and maintenance phases. Addressing the challenges of adopting digital solutions highlights the importance of developing comprehensive strategies for sustainable agricultural practices.
The advent of Industry 4.0 offers not only increased efficiency and productivity in agriculture but also a positive impact on the preservation of natural resources (Zambon et al., 2019). Tools such as soil sensors, smart irrigation systems, and Big Data enable farmers to reduce the use of water, fertilizers, pesticides, and other chemicals. These technological tools provide real-time information, allowing farmers to analyze data and make informed decisions, which in turn helps mitigate the effects of adverse climatic factors.
This study is guided by the following research question: How can AI be used in sustainable agriculture? The primary objectives are to analyze how AI optimizes agricultural resources and minimizes environmental impact within the context of sustainable agriculture. To achieve this, the research employs a mixed-methods approach, integrating structural equation modeling (SEM), multivariate regression analysis, text mining, and scenario analysis to provide a comprehensive evaluation of AI adoption. The study examines current trends and applications of AI-driven technologies in agriculture, including climate-smart agriculture, site-specific management, remote sensing, IoT, and machine learning-based decision support systems. Additionally, it seeks to identify digital gaps in sustainable agriculture and propose strategies to overcome the barriers that hinder AI adoption. A specific focus is placed on Costa Rica, assessing how AI and other digital technologies contribute to improving sustainability, productivity, and resilience in the country’s agricultural sector. By analyzing economic, environmental, and behavioral factors influencing AI adoption, the study aims to provide actionable insights for policymakers, researchers, and industry stakeholders to drive the successful integration of AI in sustainable farming practices.
2. Methodology
This research employs a mixed-methods approach, integrating qualitative and quantitative-descriptive techniques to analyze AI adoption in sustainable agriculture. The qualitative portion includes in-depth interviews with technology experts who have worked with AI in agriculture, exploring their experiences, perspectives, and perceived challenges. The quantitative portion consists of structured survey (Appendix) to gather data on technological trends and existing gaps, facilitating a comprehensive understanding of AI's role in enhancing agricultural sustainability. The study aims to detect patterns and develop strategies that promote AI integration, ensuring the resilience and efficiency of agricultural practices worldwide.
A qualitative approach, complemented by advanced statistical analysis and data visualization, is ideal for exploring complex phenomena, particularly due to the novelty of this research topic. Qualitative methods allow for an in-depth understanding of participants' experiences, motivations, and perceptions, which are often overlooked in purely quantitative research (Kim & Bradway, 2017). Descriptive statistics provide a structured way to summarize and present data, identifying patterns and trends (George & Mallery, 2018). To enhance analytical depth, this study employs multivariate regression analysis, structural equation modelling (SEM), text mining, and scenario analysis, which allow for a more nuanced exploration of AI adoption in agriculture. These advanced techniques facilitate robust findings, making the results more generalizable while retaining contextual relevance.
Primary data for this research is collected through two main sources. The first involves documentary analysis, including peer-reviewed journal articles, academic books, news reports, and relevant websites focused on AI applications in sustainable agriculture. The second source is a survey administered to IT professionals involved with AI technologies, capturing their perspectives on various AI methodologies and their potential applications in agriculture. The survey incorporates text mining techniques, allowing for topic modelling of farmers' concerns regarding AI, which provides deeper insights into sentiment and thematic trends.
The study population consists of IT professionals engaged in AI-related agricultural projects. These individuals are selected based on their expertise and experience, ensuring the reliability and relevance of their insights. According to the College of Professionals for the Field of Informatics, 11,711 registered IT professionals are currently active. To determine the sample size, factors such as confidence level and margin of error are considered. The sample size is calculated using the following equation:
Sample size = C2 Z2 × P × (1−P)
C2
Where Z represents the confidence level, P = 0.5, and C represents the margin of error. For this study, a 95% confidence level and a 5% margin of error are applied, resulting in a required sample size of 373 surveys. However, 412 responses were received, of which 383 were deemed valid based on relevance filters.
To ensure comprehensive data analysis, both qualitative and quantitative data processing methods are employed. Qualitative analysis is conducted through survey responses and in-depth interviews, utilizing sentiment analysis and topic modelling to classify common themes and concerns (Taguchi, 2018). Quantitative analysis includes multivariate regression, structural equation modelling (SEM), and scenario analysis to explore key relationships between AI adoption, economic benefits, and implementation barriers.
To facilitate data interpretation, Power BI and Python-based analytical tools are utilized for advanced data visualization. This study includes heatmaps (to illustrate AI technology adoption rates), network diagrams (to map adoption barriers and influencing factors), radar charts (to visualize AI’s economic and environmental benefits), and scenario-based projections (to estimate future AI adoption growth under different policy conditions). These visual representations enhance the analytical rigor of the study, making complex relationships more accessible and actionable.
This methodological framework provides a holistic understanding of AI's potential in sustainable agriculture. By integrating qualitative insights with quantitative data and advanced analytics, the study delivers more robust and actionable findings. This comprehensive approach supports informed policy recommendations and strategic AI implementations to drive agricultural sustainability. As illustrated in Figure 1, the research followed a sequential mixed-methods approach integrating structured surveys, statistical modeling, text mining, and scenario-based analysis to evaluate the role of AI in sustainable agriculture.
Figure 1 summarizes the integrated methodological approach used in the study. It begins with a structured survey administered to 412 agricultural professionals across Costa Rica, followed by quantitative analysis using Structural Equation Modeling (SEM), regression analysis, and descriptive statistics. Parallel qualitative analysis was conducted through text mining of open-ended responses. Scenario simulations were developed using Power BI to visualize AI adoption pathways. Analytical tools included Python (v3.11), AMOS (v29), and Power BI (April 2024 version).
Figure 1. Research Methodology Framework.
3. Results and discussion
The study explores the current landscape of AI in sustainable agriculture based on 384 responses from professionals in the field. The data highlights the nascent stage of AI adoption in agriculture, with 61% of participants having less than one year of experience using technology in farming, 29% having one to three years of experience, and only 1% having over five years. This underscores the need for more skilled professionals in AI-driven agricultural solutions.
Adoption of AI technologies in agriculture
As shown in Figure 2, perception scores across AI tools varied significantly by company size and geographic region, with smaller organizations in the Central region expressing the highest levels of familiarity and optimism. Thus, illustrating the dominance of Machine Learning (75%), followed by Deep Learning (64%), Natural Language Processing (41%), and Expert Systems (13%). The widespread use of Machine Learning aligns with its capability to analyze large datasets and improve decision-making. The preference for open-source AI platforms is also notable, with PyTorch (51%) and TensorFlow (46%) being the most commonly used tools (Dop, 2020). The data analysis reveals a strong tendency among experts toward the use of Machine Learning, which is predictable as it helps farmers extract data generated through the IoT. According to Hansen (2002), data interpretation can provide predictions about weather patterns, offering essential information for farmers to make informed decisions (Okot et al., 2023).