Available online at [Link].

org

Emerging Science Journal

(ISSN: 2610-9182)

Vol. 8, No. 2, April, 2024

Review Article

Agriculture 5.0 and Explainable AI for Smart Agriculture:

A Scoping Review

Siti Fatimah Abdul Razak 1* , Sumendra Yogarayan 1 , Md Shohel Sayeed 1 ,

Muhammad Izzat Faiz Mohd Derafi 1
1
Faculty of Information Science and Technology, Multimedia University, Melaka 75450, Malaysia.

Keywords:
Abstract
Explainable AI (xAI);
The visionary paradigm of Agriculture 5.0 integrates Industry 4.0 principles into agricultural Agriculture 5.0;
practices. Our scoping review explores the landscape of Agriculture 5.0, emphasizing the pivotal Data-Driven Agriculture;
role of Explainable AI (XAI) in shaping this domain. Guided by the Preferred Reporting Items
Smart Agriculture.
for Systematic Review and Meta-Analysis Scoping Review, we rigorously analyzed 84 articles
published from 2018 to September 2023. Our findings highlight XAI’s potential within
Article History:
Agriculture 5.0, recognizing its influence on intelligent farming. We propose a conceptual
framework for integrating XAI, emphasizing its impact on model transparency and user trust. Received: 06 November 2023
Despite transformative applications, existing literature often lacks XAI discussions. Our objective
Revised: 18 March 2024
is to bridge this gap and provide a reference for academics, practitioners, policymakers, and
educators in the field of smart agriculture that is both environmentally friendly and Accepted: 22 March 2024
technologically advanced. Published: 01 April 2024

1- Introduction
Agriculture has always been at the heart of human survival and prosperity, and its evolution has closely paralleled
the advancements in industrial paradigms. From the dawn of mechanization during Industry 1.0 to the era of
automation and data-driven decision-making in Industry 4.0, each industrial revolution has significantly impacted
agriculture [1]. The adoption of electronic technology and major IT advancements in the Agriculture 3.0 phase have
evolved into a more integrated approach with the introduction of precision farming approaches in the Agriculture 4.0
phase. Nonetheless, it is critical to recognize that the agricultural industry has specific constraints, such as a lack of
experience and IT infrastructure, that impede the rapid adoption of technolo gical breakthroughs. For example, space
technologies are of utmost importance in enhancing soil quality, mitigating water wastage in irrigation practices, and
facilitating the dissemination of agricultural knowledge among farmers. The collection, analysis, and use of geo-
spatial data from many sources, including terrestrial, aquatic, and aerial sensors, as well as satellites and surveillance
equipment, are imperative to implementing smart agriculture techniques [2], which include science, innovation, and
space technology, to enhance the production and quality of agricultural outputs [3]. The agricultural and industrial
revolutions are illustrated in Figure 1.
The historical progression of technological advancements in agriculture is examined by tracing their development
from conventional food systems to Agriculture 5.0, which represents the post-industrial era characterized by the
integration of robots, big data, and artificial intelligence systems [4–6]. Researchers have investigated the progression

* CONTACT: [Link]@[Link]
DOI: [Link]
© 2024 by the authors. Licensee ESJ, Italy. This is an open access article under the terms and conditions of the Creative
Commons Attribution (CC-BY) license ([Link]

Page | 744
 Emerging Science Journal | Vol. 8, No. 2

of industrial agriculture, considering the influence of the industrial revolution, i.e., the correlation between Agriculture
4.0 and Industry 4.0, specifically focusing on the enhancement of field work mechanization and machinery [7].
Moreover, developing countries require immediate implementation of data-driven technological advancements to boost
their gross domestic product (GDP) and guarantee food security for their population [2]. With advancements in
information technology, agricultural data analytics offers valuable and pertinent insights to farmers in the realm of
intelligent agriculture, leading to enhanced yield security and significant improvements in crop output [8]. Additionally,
these advancements will aid in conserving resources such as fertilizers, labor, seeds, and water [9], ultimately
contributing to the stability of the food supply and alleviating poverty [10, 11]. Hence, timely use of automated and
advanced technology such as the integration of drones, the use of precision technologies and equipment, the
implementation of wireless sensor networks, the development of intelligent solar and wind energy systems, the
application of artificial intelligence (AI)-based technology, and the adoption of large-scale desalination technology will
undoubtedly be the future of smart agriculture. These technologies must possess predictive accuracy, user-friendliness
(and interpretability), and the capability to facilitate decision-making that can be acted upon [12, 13]. To adequately
address the issues in smart agriculture, it is essential to comprehend the agricultural ecosystems.

Figure 1. Agriculture and Industrial Revolution

Artificial intelligence (AI) and its subsets, i.e., machine learning and deep learning algorithms, have emerged as very
influential and revolutionary advanced analytics tools that organizations may employ for making decisions [11].
Nevertheless, the accelerated progress of technology has increased the complexity and accuracy required to explain the
principles of artificial intelligence (AI) and machine learning (ML) [14]. Numerous AI models that have achieved
success are characterized by their opaqueness, rendering the comprehension of their internal mechanisms and decision-
making procedures taxing. Hence, regulatory bodies are implementing new legislation aimed at mitigating the potential
dangers pertaining to fairness, accountability, and transparency that are inherent in automated decision-making systems.
Nonetheless, the use of black box models remains debatable due to their inherent transparency and lack of interpretability
[15]. Only AI developers and those with specialized knowledge are able to comprehend the output of AI models, which
constrains the potential of AI models in providing meaningful insights based on the input data and prediction as well as
supporting decision-making processes involving different types of users [16].
Fernandez et al. [17] and Samek et al. [18] emphasize that Explainable Artificial Intelligence (XAI) involves a
selection of technologies designed to reveal the working principles of black-box models and provide explanations for
individual predictions. Compared to the word interpretability, explainability is more significant to present relevance,
comprehensiveness, and accuracy of predictions that are produced by AI models or intelligent systems. Hence, XAI is
more straightforward with a logical and transparent computational process to facilitate users in making sense of the
patterns or relationships revealed by AI models [16, 19]. According to Montavon et al. [20] and Guidotti et al. [21], the
outputs from the XAI models are more trustworthy to users since they can be easily understood, embody the principle
of transparency [22], promote accountability according to ethical principles [23], and demonstrate efficacy in mitigating
erroneous and biased decision-making [24].
Hence, the objective of this review article is to examine the potential Explainable Artificial Intelligence (XAI) in the
context of smart agriculture and specifically Agriculture 5.0. This review contributes to the existing literature as follows:
(i) Provide the focus or theme of earlier review studies between 2018-2023 related to smart agriculture, along with a
brief description of the contribution of the reviews.
(ii) Propose a conceptual framework for the integration of XAI with Agriculture 5.0 based on the fundamentals of XAI
and Agriculture 5.0.
(iii) Systematically analyze the literature on proposed XAI approaches for agriculture-related tasks using PRISMA-
ScR and discuss the challenges of implementing XAI in the context of Agriculture 5.0.
The following section presents the review methodology, which is based on the Preferred Reporting Items for
Systematic Review and Meta-Analysis Scoping Review (PRISMA-ScR) guidelines. Section 3 presents the results and
discussions based on this review, while Section 4 presents the challenges of integrating XAI in the Agriculture 5.0
landscape. The last section concludes this review paper.

Page | 745
 Emerging Science Journal | Vol. 8, No. 2

2- Methodology
The scoping review followed the PRISMA-ScR guidelines [25] to guide the search, selection, and analysis of articles.
The research question guiding this review was: "What is the current state of research on the integration of Agriculture
5.0 and Explainable AI for smart agriculture, and what are the challenges?"
To identify relevant studies, an electronic search was conducted over the Scopus database for articles published within
2018 to September 2023 to ensure the recency of the scoping review. We applied a combination of search terms related
to agriculture, artificial intelligence, and explainability. The Boolean operators AND and OR were used to refine the
search queries. The search terms used are listed in Table 1.

Table 1. Search terms used for the literature search

Agriculture farm*, digital, smart, agriculture 5.0, plant, crop, agricultur*, precision farming, precision agriculture
Explainable XAI, explanation, explainability, interpret*, trust*, ethic*, causal*, understand*
AI, machine learning, machine intelligence, deep learning, transfer learning, neural network, convolutional, black
Artificial Intelligence
box, classifier, prediction model, supervised learning, reinforcement learning, unsupervised learning, LIME, SHAP

The inclusion criteria specified for this scoping review include studies published in English as review papers, surveys,
journal articles, and conference proceedings that are related to smart agriculture and explainable AI, focusing on the
implementations, applications, challenges, and integration of both domains. The retrieved articles were independently
screened by the authors for eligibility by analyzing the title and abstract based on the inclusion criteria. Non-relevant
papers were omitted from this review. Discrepancies were resolved through discussion or by consulting the authors of
this paper. Subsequently, a comprehensive evaluation was conducted by examining the complete text of the articles to
exclude those that were not pertinent to this review, those that were excessively concise or generic, articles that did not
sufficiently address the subject of agriculture and explainability, as well as those that lacked an easily accessible full
text. Figure 2 illustrates the PRISMA-ScR flow diagram of the scoping review process.

Figure 2. PRISMA-ScR flow diagram

Furthermore, relevant information was extracted from the potential studies, i.e., study objectives, methodology, key
findings related to smart agriculture and XAI, as well as the implications for users. The data that was obtained was
synthesized and thereafter presented through the use of descriptive and thematic analysis techniques. The present study

Page | 746
 Emerging Science Journal | Vol. 8, No. 2

identified and provided a summary of the key themes and patterns pertaining to the integration of Agriculture 5.0 and
XAI in the context of smart agriculture. Furthermore, the objective of the scoping review was to offer a comprehensive
summary of the existing body of research in this field. The present research demonstrates that there has been an
increasing trend in the interest and attention given to the agricultural applications of artificial intelligence that employ
approaches for explainability. Moreover, it is important to highlight that there is a clear focus on explaining the
techniques used in deep learning models, considering their current dominance as the most often applied machine learning
methods.

3- Results and Discussion

3-1- Agriculture 5.0 in Smart Agriculture
The advent of Agriculture 5.0 represents a unique and transformative phase. Basically, the need for smart agriculture
arises from the need to address challenges such as population growth, resource scarcity, climate change, cost
optimization, and labor efficiency. Resources such as water, land, and energy are limited and face increasing pressure
due to population growth and climate change. Smart agriculture has the capacity to revolutionize conventional farming
methods by incorporating cutting-edge technologies, resulting in more sustainable, efficient, and eco-friendly systems
[26]. This fusion of cutting-edge technologies has ushered in a new era of precision farming, often referred to as smart
agriculture or digital agriculture. By incorporating geographical and temporal variations into technological
advancements and the data-driven decision-making process, smart agriculture facilitates accurate resource management,
minimizes waste, and maximizes the efficient use of resources such as water, fertilizers, and energy [27, 28]. This
eventually enhances economic efficiency while simultaneously minimizing resource inputs and environmental
consequences [29–31]. Farmers will be empowered to make data-driven decisions based on events [32] at different stages
of production, such as extreme weather, changing rainfall patterns, and the spread of pests and diseases, through
optimized resource usage [33, 34] by utilizing robotics, automated control systems, and artificial intelligence methods
[35], as well as integrating agricultural research and information to optimize the productivity and efficacy of the
agriculture industry [36, 37]. Thus, farmers do not need to rely on manual labor for tasks such as planting, weeding, and
harvesting, which enhances worker safety by minimizing exposure to hazardous conditions [38]. It also emphasizes
collaboration and cooperation between humans and machines. The integration of robots and AI algorithms into data-
driven farms, which enable seamless interaction between farmers, autonomous machinery, and data-driven algorithms,
is known as Agriculture 5.0 [5, 9, 39]. The elements of Agriculture 5.0 are summarized in Table 2, and the concept of
Agriculture 5.0 adopted from Polymeni et al. [40] is illustrated in Figure 3.

Table 2. Elements of Agriculture 5.0

Elements Description
Agriculture 5.0 flourishes within an environment with extensive connectivity between devices and streams of data,
Connectivity and wherein gadgets, machines, sensors, and instantaneous data streams. The advent of 5G technology, coupled with the
Data ubiquity Internet of Things (IoT), entails the widespread presence of data, leading decision-makers to possess a thorough and
current comprehension of their agricultural systems.
Data-driven decision Optimize agriculture operations or processes by employing a fusion of advanced analytics, machine learning algorithms,
making and predictive modeling based on a thorough understanding of the data environment.
Robotics and Minimize harmful interventions such as chemical usage to promote and foster harmonious coexistence of human expertise
Automation and machine precision in executing tasks with precision and efficiency.
Precision ecosystem Ensure sustainable agricultural practices that hinder the environmental ecosystem from being compromised to provide
and Sustainable lasting resources, including productive soils, uncontaminated water supplies, and flourishing ecosystems for future
Practices generations.

Figure 3. Agriculture 5.0 concepts

Smart agriculture provides a means to enhance agricultural output and efficiency to satisfy the increasing demand of
estimated 9.7 billion people by 2050 [5]. By the year 2100, it is projected that over ten billion individuals would require
food systems that exhibit carbon neutrality, resilience to catastrophic weather events, and the ability to provide culturally
suitable nourishment in a financially viable manner, all while adhering to the limitations imposed by the planet. The

Page | 747
 Emerging Science Journal | Vol. 8, No. 2

resolution of this worldwide dilemma requires the implementation of a novel agricultural revolution that facilitates the
integration of technology aimed at enhancing productivity while concurrently ensuring environmental sustainability.
There is a pressing need for multidisciplinary and practical research and synthesis pertaining to food and farming systems
to address the challenges faced by humanity. Similarly, it is necessary to use these understandings in manners that are
culturally, economically, and politically suitable to guarantee the production and distribution of food in a manner that
surpasses the current levels of economic and ecological efficiency [4].
Agriculture 4.0 is founded upon four fundamental pillars, namely data-driven management, the utilization of novel
tool-based production methods, a commitment to sustainability, and the pursuit of professionalization [41]. The
emergence of Agriculture 5.0 can be understood as a natural extension of the Agriculture 4.0 and Industry 4.0 principles
into the realm of agriculture. Furthermore, the agricultural sector in European nations is now experiencing significant
growth after the accessibility of GPS signals [42]. The integration of Agriculture 5.0 into the strategic plans of prominent
agricultural equipment manufacturers is a prominent objective for the forthcoming decade, as highlighted by Saiz-Rubio
and Rovira-Más [5]. Agriculture 5.0 is the envisioned paradigm shift in the future of smart agricultural practices. During
that period, the prevalence of unmanned farming is expected to increase, hence offering advantages in situations where
there is a scarcity of labor.
Agriculture 5.0 is more likely to produce more food with fewer inputs and less land with the implementation of
cutting-edge digital technologies and biotechnologies [4]. Farm-to-fork traceability will be improved [26], allowing
farmers to track the entire supply chain, ensure food safety, and provide transparent information to consumers [43, 44].
Moreover, studies on Agriculture 5.0 and food security observe that the application of artificial intelligence and robotics,
along with machine learning (ML) and deep learning (DL), has the potential to forecast crop yields with a 75%
probability [45, 46]. However, even though agriculture practices should increase productivity, the environment should
not be compromised [5]. Strategies should be crafted to focus on gender empowerment, market accessibility, utilization
of low-cost tools, minimizing food waste, and effective food distribution [4]. Hence, it is pertinent to contemplate the
concept of Agriculture 5.0 in correlation with Industry 4.0 and Society 5.0, with the objective of advancing ecologically
responsible technologies, establishing resilient smart cities, and safeguarding industrial progress as a fundamental
measure of human welfare [47]. Moreover, the use of Agriculture 5.0 technology on the farm is important to facilitate
precision farming and enable the realization of its full potential in sustainable global food systems [4]. At the core of this
transformational process lies the incorporation of Explainable Artificial Intelligence (XAI), a technological advancement
that not only equips farmers with sophisticated decision-making assistance but also guarantees the ethical foundations
of AI-based agricultural systems. Table 3 summarizes reviews or surveys that have been published in relation to smart
agriculture.
Table 3. Smart agriculture-related reviews or surveys

Author (year) Focus Descriptions Remarks

 Investigates the implementation of supervised and unsupervised
ensemble and deep learning techniques in predicting yield,
irrigation, and environmental elements for smart farming.
Kwaghtyo et  Poor model performance is attributed to dataset quality and  Narrative review
General
al. (2023) [48] preprocessing issues.  No reference to XAI
 The article suggests a machine learning procedure to mitigate
dataset quality and preprocessing issues and emphasizes the
potential of automation in agriculture.
 Highlights security concerns in smart agriculture, including
Otieno (2023) unauthorized access and operational disruption, which are  Narrative review
Security
[26] attributed to IoT devices and data communication, as well as  No reference to XAI
farmers’ security awareness to protect their data.
 Introduces the scope of Agriculture 5.0 and its impact on Quality
of Service, security, and reliability in the 6G-IoT landscape.
Polymeni et al. Agriculture 5.0  Narrative review
 Illustrates the potential of combining existing precision algorithm
(2023) [40] and 6G-IoT  No reference to XAI
concepts with quantum sensing, robots, enhanced wireless
networks, and artificial intelligence for Agriculture 5.0.

 Summarizes the current state in the field of automatic plant disease

detection and indicates that CNN models, particularly VGG
models, exhibit superior performance in the domain of plant
disease detection. Accuracy is the most employed performance  Adopted tertiary systematic literature
Teeffelen Plant Disease metric. review methodology.
(2023) [49] Detection
 Develop a VGG16 transfer learning model to analyze the  No reference to XAI
influence of datasets on the outcomes of plant disease detection
models. Three distinct datasets with varying features (small/large,
in-situ/laboratory, raw/processed) are employed.

Page | 748
 Emerging Science Journal | Vol. 8, No. 2

 Analyzed articles from 2010–2020 that have applied machine

learning to crop disease before presenting a framework for the
future of precision crop protection, emphasizing the scientific,
Mesias-Ruiz et agronomic, and industrial uses of traditional machine learning  Narrative review
Crop Protection methods and current advancements in artificial neural network
al. (2023) [6]  No reference to XAI
models.
 The aim of the framework is to advance precision crop protection
and align it with the emerging concept of Agriculture 5.0.

 Analyzed and explored the obstacles and potential advantages of

enhancing farmers’ confidence in AI technologies for precision  Narrative review
Gardezi et al. agriculture.  Partly mentioned XAI to consider
Trust
(2023) [50]  Emphasized the necessity of regulating and ensuring voluntary farmers training needs when engaging
adherence to data ownership, privacy, and security to establish the with XAI
acceptance and utilization of AI systems among farmers.

 Reviewed smart irrigation monitoring and control strategies to  Narrative review

Bwambale et
Irrigation System enhance water use efficiency in precision agriculture, including
al. (2022) [29]
irrigation scheduling and monitoring the weather, soil, and plants.  No reference to XAI

 Provided a brief overview of smart agriculture, focusing on  Narrative review

Vikrant et al.
General Wireless Sensor Networks (WSN), IoT, and data analytics
(2021) [51]
applications, benefits, challenges, and directions.  No reference to XAI

 Described the use of intelligent sensors, IoT sensor types,

networks, and protocols, as well as IoT applications and solutions
Boursianis et in smart farming.  Narrative review
IoT and UAV
al. (2022) [52]  Highlighted the role of UAV technology in irrigation, fertilization,  No reference to XAI
use of pesticides, weed management, plant growth monitoring,
crop disease management, and field-level phenotyping.

 Focused on fairness, transparency, accountability, sustainability,

privacy, and robustness of AI in agriculture with the aims to  Narrative review
[53] Dara et al.
Ethics address farmers’ privacy concerns, ensure reliable AI
(2022) [53]
performance, enhance sustainability in AI systems, and reduce AI  No reference to XAI
bias.
 Adopted the three-stage systematic
 Presented a systematic review of 93 research papers discussing
literature review (SLR) methodology,
Sharma et al. machine learning (ML) applications in agricultural supply chains
Supply Chain comprising of the pre-operational,
(2022) [11] (ASCs) and highlighted the benefits towards sustainable ACS.
operational and post-operational stage.
 Proposed an ML application framework for sustainable ASC.
 No reference to XAI
 Provides a survey of EU research projects and scientific literature
in smart farming and categorizes them into four categories, i.e.,
sensing techniques and management systems, unmanned vehicles,
Bacco et al. IoT platforms, and decision support systems.  Narrative review
General
(2019) [54]  Identify threats and concerns in non-technical (incentives,  No reference to XAI
investments, and innovative tools) and technical (data, networks,
and information), and then look at existing and upcoming
solutions to overcome those barriers in the EU research project.

 Focuses on artificial neural networks, automation, wireless system  Narrative review

Jha et al. Artificial
networks, and the implementation of fuzzy logic in agriculture for
(2019) [8] Intelligence
flower and leaf identification and watering using IOT.  No reference to XAI

 Identify the relationship between digital architecture and farm  No reference to XAI
Klerkx et al.
(2019) [55]
Social Science diversity and examine the impact on business practices, value  Perform exploratory literature review
chain, food system, economy, and innovation system. with thematic clusters

3-2- XAI Methods

Modern technologies that employ deep learning, big data, and IoT architectures necessitate the extensive use of
complicated computing technologies, which are normally known as black box models. User access is limited and
depends on the tools employed. The comprehensibility of the outcomes generated by the models is also lacking, which
presents a substantial obstacle to the adoption of AI-based decisions. In this context, Explainable Artificial Intelligence
(XAI) is used to clarify these decisions [56] by converting complex systems into clear systems, allowing for
understandable, interpretable, and transparent decision-making processes [57]. The value of human expertise should not
be underestimated but rather complimented by AI, which utilizes formal modeling and reasoning approaches to solve
real-world problems [56].
XAI techniques can be categorized into the explanation scope, which can either be local, focusing on specific
predictions made by the model, or global, aiming to characterize the overall behavior of the model. Comprehensive
explanations enhance comprehension of the overall behavior and logic of the model, hence leading to anticipated

Page | 749
 Emerging Science Journal | Vol. 8, No. 2

consequences. Local explanations refer to the provision of explicit justifications for a single prediction, explaining the
rationale for the model's judgment for that case [16]. They can also be categorized based on their applicability to any
machine learning algorithm (model agnostic) or their limited applicability to a specific machine learning algorithm
(model specific). Model-specific techniques have specific constraints for the type of data or procedure employed,
whereas model-agnostic methods are more flexible and can be applied in any situation [22].
Another categorization is based on the type of explainability model, either an intrinsic interpretable model or a post-
hoc model. Intrinsic interpretable models, or transparent, are also known as white box/glass box approaches. The model's
architecture and decision-making process are intentionally designed to facilitate human comprehension of its prediction
methodology. These models are specifically designed using logical relations, statistical, or probabilistic frameworks to
offer transparency and interpretability without requiring any supplementary post-hoc interpretation techniques. For
instance, linear and logistic regression, decision trees, k-nearest neighbors, fuzzy inference systems, rule-based learners,
general additive models, and Bayesian models fall under this category [16], with a similar perspective taken by
knowledge-based systems. Nevertheless, their level of explainability, transparency, or interpretability is limited when
high-dimensional scenarios are involved [58]. On the other hand, post-hoc models are additional features implemented
on results or completed black-box models like random forests, support vector machines (SVMs), tree ensembles, and
deep learning methods [22]. They are called post-hoc methods because they generate post-hoc explanations that are
independent of the model used, in contrast to the approach of ante-hoc models. However, some of these methods may
have certain requirements related to the data or structure of the model. Additionally, post-hoc procedures can be
evaluated using quantitatively quantifiable metrics and human-centered evaluations [59]. Nevertheless, there is presently
a lack of agreement over the criteria for assessing the interpretability of a model, the accuracy of an explanation, and the
methodologies for comparing different approaches. Moreover, post-hoc explanations are subject to concerns over their
dependability [60].
Further categorizations can be established based on the way the output explanations are presented (such as textual,
visual, or rule-based), the specific input data that is necessary, the problems to which they might be applied, or the
method by which they are generated. According to Nauta et al. [61], XAI methods can be evaluated based on goodness
checklists, explanation satisfaction scales, elicitation methods for mental models, computational measures for explainer
fidelity, explanation trustworthiness, and model reliability. Figure 4 illustrates the most common approaches to the
classification of XAI methods.

Figure 4. Common classification of XAI methods

3-3- Integration of XAI in Agriculture 5.0

Recent years have shown that XAI has gained attention for improving smart agriculture processes, including
enhancing crop production, minimizing resource waste, reducing environmental impact, and improving overall
profitability. This is attributed to the vast amounts of data acquired through sensors, drones, and satellite imagery [2,

Page | 750
 Emerging Science Journal | Vol. 8, No. 2

62]. Analyzing this data using AI and data analytics helps farmers gain valuable insights, make informed decisions, and
optimize farming practices for better outcomes [63]. For instance, data generated from continuous monitoring of crops,
including factors like soil moisture, temperature, and nutrient levels, may help farmers detect crop stress, diseases, and
nutrient deficiencies early on, allowing timely interventions and reducing crop losses [64, 65]. Based on the literature,
XAI can be integrated into Agriculture 5.0 and XAI for enhancing decisions, enabling autonomous farming operations,
and localizing decision-making [26]. It is an innovative approach to agricultural practices that leverages advanced
technologies to optimize efficiency, productivity, and sustainability in farming operations [41, 48, 52, 66], leading to
innovative applications that fulfill the basic requirements of Agriculture 5.0 [40]. A conceptual framework for the
integration of XAI in smart agriculture is proposed and illustrated in Figure 5.

Figure 5. Agriculture 5.0 and XAI

Based on Figure 5, XAI requires a collection of diverse data sets from sensors or IoT devices in real-time using the
Internet of Things (IoT), artificial intelligence (AI), data analytics, robotics, and remote sensing to create a
comprehensive dataset for analysis. By utilizing a network of interconnected devices and sensors, farmers can collect
real-time data on various factors, such as soil moisture levels, weather conditions, crop health, and pest infestations. This
data is then analyzed to provide actionable insights, allowing farmers to make data-driven decisions and implement
precise interventions [11, 51, 54, 55, 67].
The data is generally integrated into a platform for data cleaning and pre-processing to take place. Depending on the
tasks at hand, feature engineering can be conducted to extract relevant information and enhance model performance [5].
Next, machine learning models will be trained for various agricultural tasks, including fertilizer application, weed or
pest control, yield mapping, seed sowing, irrigation, etc., depending on the need. The XAI developer must ensure that
the models are interpretable and explainable, even if they are complex. Moreover, the interpretability of the
recommendations, which are supported by implementing visualizations and summaries from XAI techniques, may
improve the transparency and accountability of decision-making. The model's performance will be assessed based on
metrics that capture both accuracy and interpretability of the model and generalization across different scenarios.
Besides, user-friendly interfaces that present AI-driven insights in a comprehensible manner to farmers and other
stakeholders are a necessity to allow users to interact with the AI system, ask questions, and receive explanations for
model predictions. Farmers and stakeholders will acquire a clear understanding of how AI-driven decisions are made.
Eventually, they will be empowered to trust and harness AI-driven insights, making informed decisions that optimize
crop yields, resource utilization, and sustainability as envisioned by Agriculture 5.0. As responsible innovation, it is
essential to provide training programs for end-users to understand the basics of AI and the specific models used in
Agriculture 5.0. The training should foster knowledge transfer between AI experts and farmers to enhance collaboration
and address domain-specific challenges. There should also be mechanisms for continuous monitoring of model
performance in real-world conditions where the models can be updated based on new data and feedback from users to
ensure relevance and accuracy.
Guidelines and protocols for responsible AI use in agriculture should be developed to address ethical considerations
related to data privacy, bias, and fairness in AI models. A collaborative ecosystem involving farmers, researchers,
policymakers, and AI developers should be available to encourage knowledge sharing, feedback loops, and the co-

Page | 751
 Emerging Science Journal | Vol. 8, No. 2

creation of AI solutions that align with the needs of the agricultural community. Based on the literature, previous work
has applied XAI in the context of smart agriculture. Table 4 outlines the focus of the work, describes the goals in brief,
lists the tools or pre-trained models, provides the results of the best-performing model in the study, and classifies the
type of explanation provided.
Table 4. XAI applications in smart agriculture
Tools/ Pre-Trained Type of
Author (Year) Focus Descriptions Evaluation
Models Explanation
Combine vision models generated using deep
learning with semantic models generated using
ontologies and knowledge graphs. The fusion RESNEXT50,
Chhetri et al. Cassava
model was assessed based on prediction accuracy EfficientNetV2S, and Accuracy 90.5% Text
(2023) [68] Disease
and latency, user-level explainability and VOLO
understandability, as well as explanation
usefulness.
Propose a Bayesian ensemble model (BM) that
outperforms ElasticNet, Neural Network,
Climate-crop Accuracy 99.2%;
Hu et al. (2023) MARS, SVM, Random Forests, and XGBoost in Bayesian ensemble model
Yield Visual
[69] both predicting and explaining. BM was the only (BM) RMSE 0.50
Relationship
method for unveiling true relationships from
synthetic data.
Identify and classify plant diseases based on the YOLOv5: F1 score 0.57
severity of the disease using deep learning (conf. 0.2);
networks, and offer appropriate treatments to
Bandi et al. ViT (with background): F1
Leaf Disease stop the disease from spreading to other plant YOLO5 and VIT Text
(2023) [70] score 0.758;
leaves.
ViT (without background):
YOLOv5 is used to detect disease, and the vision
F1 score 0.908
transformer (ViT) is used for the classification.
Discuss the implementation of an XAI model
based on a Random Forest classifier to predict the
Chandra et al. relative soil fertility of a given soil using its Random Forest, Accuracy 97.02%;
Soil Fertility Visual
(2023) [71] various physiochemical properties and explain (SHAP) F1 score 0.87
the reasons behind the model’s soil fertility
indicator prediction using user-friendly graphs.
Combine ML methods, i.e., category boosting
(CatBoost), extreme gradient boosting
(XGBoost), support vector machine (SVM), k
nearest neighbor (KNN), multi-layer perceptron
Sahidullah et al. Fruit SVM, KNN, MLP,
(MLP), logistic regression (LR), and bagging F1-score 0.90 Visual
(2023) [72] Classification boosting, bagging, (LIME)
methods with Local Interpretable Model-
Agnostic Explanations (LIME), to classify
different types of date fruits. MLP outperforms
the other methods.
Compared the performance of a glass box
method known as Explainable Boosting Machine
(EBM) to the traditional regression models (RLR
and LASSO) and tree-based models (DT,
Celik et al. Cotton Yield Explainable boosting MAE 87.41, RMSE 118.57, Visual
XGBoost, and LightGBM) to predict cotton
(2023) [73] Prediction machine (EBM) MAPE 0.10, and R2 0.73
yield. The study findings showed that
precipitation (P), enhanced vegetation index
(EVI), and leaf area index (LAI) are the three
most important dynamic features.
Develop Gradient Boosted Regression Tree
(GBRT)-based deep learning surrogate models,
combined with a Bayesian optimization (BO)
Bhat et al. algorithm to determine the most optimal GBRT-based hybrid DNN
Crop Yield F1 score 1.0 Visual
(2023) [74] hyperparameters for the deep neural network. surrogate models
The impact of each input parameter on the
individual outputs is evaluated using explainable
artificial intelligence (XAI).
Compared permutation-based variable
Linear model with AIC
importance (global), pairwise interaction
stepwise variable selection,
importance (global), partial dependence plot RMSE 0.199;
Ryo (2022) [75] Crop Yield conditional inference tree, Visual
(global), and LIME local variable importance for R2 0.42
random forest, gradient
all pre-trained models. Random forest is the best
boosting
algorithm based on this approach.
Kawakura et al. Demonstrate the capabilities of XAI techniques
Workers ELI5, PDPbox and Skater Not provided Visual
(2022) [76] using wearable sensor data and questionnaires.

Page | 752
 Emerging Science Journal | Vol. 8, No. 2

Demonstrates the use of LIME and SHAP in Extreme Gradient Boosting

Accuracy:
recommending crops to cultivate based on XGB, (XGB), Multi-Layer
Cartolano et al. Crop XGB 99.32
MLP, and SVM classifiers. XGB obtained the Perceptron Neural Visual
(2022) [77] Cultivation Linear SVM: 98.64
highest accuracy with SHAP’s TreeExplainer. Network (MLP), Support
MLP: 97.05
Vector Machine (SVM)
Evaluate the suitability and applicability of Gaussian Naïve Bayes,
different black-box algorithms using real data. Quadratic Discriminant
QDA and RF:
Also, we applied XAI-based data farming output Analysis, Random Forest,
Feldkamp et al. Data Farming Accuracy 99.994%; average
analysis using permutation feature importance. Nearest Neighbor,Ada- Text
(2022) [78] Output prediction time 0.12
Quadratic Discriminant Analysis (QDA) and Boost, Artificial Neural
seconds
Random Forest are preferred for subsequent XAI Network, Support Vector
analysis of the data farming output. machine (SHAP)
Used the transfer learning approach for detecting
Mehedi et al. leaf diseases with EfficientNetV2L. The
Leaf Disease LIME Accuracy 99.63% Visual
(2022) [79] approach outperforms MobileNetV2, and
ResNet152V2.
Applied XAI to analyze time series data from
Kawakura et al. SHAP, LIME, and
Workers workers who perform specific tasks, including Not provided Visual
(2022) [80] LightGBM
acceleration and angular velocity.

Most of the work highlighted the importance of interoperability and seamless integration of various components
within the agricultural value chain. This included the integration of IoT devices, cyber-physical systems, and automated
machinery into farm operations. Since Agriculture 5.0 promotes a symbiotic relationship between farmers with advanced
technologies, the studies also stressed the need for human oversight and decision-making alongside automated processes.
The reviewed studies repeatedly highlighted the significance of transparency in decision-making processes powered by
AI in the field of agriculture [22, 69, 78, 81]. Besides that, the concept of XAI was found to be particularly relevant in
agriculture practices, where autonomous systems can make real-time decisions based on local conditions and the
principles of precision agriculture [46]. This encourages the localization of decision-making and fosters human-machine
collaboration. The utilization of models, frequently employing decision trees, rule-based systems, or local explanations,
has been seen as crucial in establishing confidence in the suggestions generated by AI [82, 83]. XAI methods aim to
make machine learning models more interpretable, allowing farmers and stakeholders to understand why certain
decisions are made. Additionally, visualization tools and dashboards that present AI-generated insights in a
comprehensible manner were commonly mentioned. These interfaces allow farmers to monitor crop health, weather
forecasts, and equipment status easily.

4- Challenges in XAI for Agriculture

The scoping review revealed that while XAI holds promise for agriculture, it is not without challenges. Agriculture
5.0 concepts such as interoperability, human-machine collaboration, and decentralized production, along with XAI
methodologies that prioritize openness, have the potential to lead to more efficient, sustainable, and yield-optimized
agriculture practices. However, XAI is only meaningful in certain instances when the decision made by AI can be
elucidated on an individual basis. Ultimately, this scoping review provides a fundamental reference for researchers,
practitioners, policymakers, and educators who are interested in the intersection of Agriculture 5.0 and XAI for smart
agriculture to consider tackling the related difficulties in their work.

4-1- Real-time Data

Agriculture involves diverse and complex data types, including geospatial data, sensor data, and historical records.
The fairness of decisions produced by XAI systems relies on the data provided to the machine learning algorithms. In
their work, Mehedi et al. [79] reported challenges in acquiring adequate data and infrastructure for detecting plant
diseases using pre-trained models and improving reliability using LIME. Reliable connectivity is essential to avoid signal
delay and loss in acquiring data from multiple sensors. In addition, integrating data from different types of sensors, which
involve different data requirements like format, size, and structure, is also an important challenge to address prior to
developing any predictive AI models. In addition, farmers also find it difficult to optimize crop yield, manage production
and inventory according to customer demands, and optimize their work schedule, which could lead to labor savings and
high crop production [84]. Insufficiency of available data may cause gaps in the training data, model, and objective
function of the system.
With the growing emphasis on data in agriculture, the matter of data privacy and security becomes a crucial problem.
The widespread accumulation and dissemination of delicate agricultural data, ranging from harvest quantities to soil
condition, presents difficulties with data ownership, consent, and safeguarding. Farmers and stakeholders must confront
inquiries over data ownership, storage, and sharing, as well as the possible hazards associated with data breaches.
Adhering to ethical protocols for data management and implementing strong cybersecurity measures are essential to
safeguarding the integrity and confidentiality of agricultural data.

Page | 753
 Emerging Science Journal | Vol. 8, No. 2

To explain the reasoning of a machine learning model, it is important to possess the data necessary to train the model.
Nevertheless, the data may be exclusive or classified, making it not possible to be disclosed to the users. Since there is
a possibility of exposing confidential details in the underlying data, precautions and preventive measures need to be
implemented to maintain data privacy and security. Further research on safeguarding organizations from security attacks
that could impact AI models explainability for Agriculture 5.0 should consider developing areas like edge computing,
microservices, and 5G connections.

4-2- Ethical Considerations

The understanding of XAI is highly influenced by the context in which it is being examined [85]. Dara et al. [53]
present a framework for ethical AI in agriculture that includes six measures, i.e., fairness, transparency, accountability,
sustainability, privacy, and robustness. XAI models are susceptible to biases present in the data and the algorithms
employed for analysis. Algorithmic bias refers to the occurrence of AI systems generating findings that are distorted,
ranging from showing preference for specific crops or agricultural methods to sustaining disparities among diverse
farming groups. Additionally, the functioning and training of XAI systems rely on parameters that are determined by
developers. The level of complexity of an algorithm's context directly correlates with the difficulty of XAI. Various
technologies, such as LIME, SHAP, and Grad-CAM, can be employed to elucidate the models. However, these
technologies are tailored for particular model types or datasets, and certain tools can be intricate and challenging to
operate [59]. Furthermore, these technologies suffer from scalability and coverage limitations. The selection of
algorithms can provide different outcomes for distinct individuals [86, 87].
Decisions are also made by considering a sophisticated interaction of data and perhaps several algorithms. During the
process of converting this information into a form that a wider audience can comprehend, the developer also makes
subjective choices that may be up to debate due to inherent biases. Moreover, a singular explanation would not suffice
to account for all decisions made by an algorithm. Therefore, to mitigate the bias, it is crucial to meticulously evaluate
the data sources and algorithms employed in these systems. Cartolano et al. [77] focused on improving the explainability
of single-class classification by utilizing the visualization features of SHAP and LIME. Mehedi et al. [79] used pre-
trained weights from ImageNet datasets to customize EfficientNetV2L, MobileNetV2, and ResNet152V2 to avoid
algorithm bias among the models. Moreover, inaccurate or highly biased domain knowledge can also act as a bottleneck.
Therefore, even though Chhetri et al. [68] recommend the incorporation of domain knowledge in the models, it should
be done with extreme caution, and multiple domain experts should be involved to eliminate or reduce bias.
Ensuring equity in AI-powered decision-making is not just a moral obligation but also a pragmatic requirement to
avert unexpected repercussions and inequities in agricultural results. Although Agriculture 5.0 prioritizes sustainability
and environmental care, there are apprehensions that the extensive implementation of technology could unintentionally
result in environmental harm. For example, the utilization of self-governing machinery and unmanned aerial vehicles
might result in environmental consequences, and over dependence on technology may lead to the erosion of conventional
farming methods. Ethical problems also include the displacement of rural labor caused by automation and the possible
consolidation of power among technology providers.

4-3- Collaboration and Responsible Innovation

The responsible adoption of Agriculture 5.0 with self-governing systems and decision support powered by
artificial intelligence requires well-defined boundaries of authority and accountability in case of AI faults,
malfunctions, or misinterpretations of suggestions from AI models. Collaboration efforts between academia, industry
stakeholders, and government in building appropriate frameworks for interdisciplinary cooperation that promote
ethical deployment of AI in Agriculture 5.0 need to be strategized effectively. All stakeholders must anticipate and
prepare for the potential implications of XAI on the decision-making process in the context of Agriculture 5.0. For
instance, how they affect labor arrangements, knowledge production, and the distribution of benefits and workload
among farmers. The stakeholders must be able to adapt and respond to new information or changes in context as
agricultural systems evolve with the integration of AI and robotics [5]. This is supported by the meta-analysis by
Velten et al. [88], which underscores the success factors for collaborative groups, emphasizing the need to define
priorities, address trade-offs, and recognize agency within collaboratives. Furthermore, Maryono et al. [89] reviewed
multi-stakeholder partnerships in agricultural research projects, emphasizing the complexity of participatory research
and the importance of involvement across research contexts and phases. Hermans et al. [90] recognize the role of
social networks and effective stakeholder communication in agricultural innovation and achieving sustainability
goals. It is challenging to elucidate the functionality of the algorithms in a comprehensive manne r and to clarify their
distinct results. Different stakeholders may have diverse expectations and demands about the concept of
explainability [91].
Responsible innovation must take precedence, ensuring that technology is utilized to the advantage of everyone, with
a particular emphasis on fair access and sustainability. This includes efficient use of resources, reducing pollution, and
the conservation of ethical farming practices [5] while minimizing negative impacts on natural resources [92, 93].

Page | 754
 Emerging Science Journal | Vol. 8, No. 2

It involves designing technologies that align with farmers’ values, cultural practices, and environmental stewardship.
Key dimensions of responsible innovation include anticipation, inclusion, reflexivity, and responsiveness. For instance,
Chhetri et al. [68] employ the microservices architecture to deploy the vision model, semantic model, and decision engine
for the reusability of cassava disease classification. The BentoML and the semantic classifier REST (REpresentational
State Transfer) Service were utilized for this purpose.

4-4- Decision Support

The synergy between XAI techniques and real-time data from IoT devices and sensors, facilitated by Agriculture 5.0,
ensures that the decisions made by AI models are understandable and justifiable by the users [94]. The combination of
XAI techniques and real-time data from IoT devices and sensors enhances the capacity of farmers to harness AI-driven
insights effectively, thereby optimizing resource management and crop yields. Analyzing the data can help in providing
better services, predicting trends, and timely decision-making for industries [95].
Besides, it is vital to prioritize work aimed at enhancing the accessibility and affordability of AI and XAI for all types
of farmers to ensure fairness throughout the agriculture industry. Generally, XAI operates under the assumption that the
general user possesses a particular level of expertise. Nevertheless, it is important to acknowledge that many individuals
may lack this expertise and, hence, may not be able to accurately evaluate the fairness or justice of a decision made by
an AI system. Certain users may lack the requisite foundational knowledge to completely grasp the supplied explanations
[58]. Hence, XAI systems must be specifically designed to offer explanations that are customized to the requirements
and levels of knowledge of various user groups. This may involve incorporating visual aids, interactive interfaces, or
other tools to aid users in comprehending the rationale behind AI decisions [84, 96]. Furthermore, AI models are always
being improved to meet the changing needs of stakeholders. Complex machine learning models that involve a multitude
of parameters and complex structures, such as deep neural networks, require more calculations as the explanations must
scrutinize and decipher the intricate linkages and patterns acquired by the model. Hence, it causes high computational
costs, which may limit the applicability of XAI in the actual environment. On the other hand, in order to create models
that are easier to understand, there is frequently a trade-off that requires sacrificing accuracy [58, 88, 89].

5- Conclusion
The scoping review analyzes 84 articles from 2018 to September 2023 which offers insights on XAI’s potential impact
on crop production, resource conservation, and yield security for the future of farming. Key findings include the
recognition of XAI’s pivotal role within Agriculture 5.0, emphasizing model transparency and user trust. By bridging
the gap in existing literature, this review provides a reference for environmentally friendly and technologically advanced
smart agriculture. The proposed conceptual framework for integrating XAI serves as a practical guide for informed
decision-making. As we move toward a data-driven agricultural landscape, the timely adoption of automated and
advanced technologies such as drones, precision equipment, wireless sensor networks, and AI-based systems will
undoubtedly shape the future of farming. These technologies must prioritize predictive accuracy, user-friendliness, and
actionable decision support. However, the challenges related to the complexity of agricultural data, scalability concerns,
educational obstacles, and the trade-off between interpretability and performance must be tackled cautiously. Overall,
this review contributes significantly to sustainable food production, resource conservation, and poverty alleviation,
positioning smart agriculture as a powerful force for global well-being. In summary, the conclusion effectively
encapsulates the review’s key findings and underscores their implications for advancing agriculture into a more efficient,
transparent, and resilient era.

6- Declarations
6-1- Author Contributions
Conceptualization, S.F.A.R.; methodology, S.F.A.R. and S.Y.; investigation, S.F.A.R., S.Y., M.S.S., and M.I.F.M.D.;
resources, M.S.S.; writing—original draft preparation, S.F.A.R. and S.Y.; writing—review and editing, S.F.A.R. and
M.S.S. All authors have read and agreed to the published version of the manuscript.

6-2- Data Availability Statement

The data presented in this study are available in this article.

6-3- Funding
The authors received no financial support for the research, authorship, and/or publication of this article.

6-4- Institutional Review Board Statement

Not applicable

6-5- Informed Consent Statement

Not applicable

Page | 755
 Emerging Science Journal | Vol. 8, No. 2

6-6- Conflicts of Interest

The authors declare that there is no conflict of interest regarding the publication of this manuscript. In addition, the
ethical issues, including plagiarism, informed consent, misconduct, data fabrication and/or falsification, double
publication and/or submission, and redundancies have been completely observed by the authors.

7- References
[1] Mazzetto, F., Gallo, R., & Sacco, P. (2020). Reflections and methodological proposals to treat the concept of “information
precision” in smart agriculture practices. Sensors, 20(10), 2847. doi:10.3390/s20102847.

[2] Goel, R. K., Yadav, C. S., Vishnoi, S., & Rastogi, R. (2021). Smart agriculture – Urgent need of the day in developing countries.
Sustainable Computing: Informatics and Systems, 30. doi:10.1016/[Link].2021.100512.

[3] Khanna, A., & Kaur, S. (2019). Evolution of Internet of Things (IoT) and its significant impact in the field of Precision Agriculture.
Computers and Electronics in Agriculture, 157, 218–231. doi:10.1016/[Link].2018.12.039.

[4] Fraser, E. D. G., & Campbell, M. (2019). Agriculture 5.0: Reconciling Production with Planetary Health. One Earth, 1(3), 278–
280. doi:10.1016/[Link].2019.10.022.

[5] Saiz-Rubio, V., & Rovira-Más, F. (2020). From smart farming towards agriculture 5.0: A review on crop data management.
Agronomy, 10(2), 207. doi:10.3390/agronomy10020207.
[6] Mesías-Ruiz, G. A., Pérez-Ortiz, M., Dorado, J., de Castro, A. I., & Peña, J. M. (2023). Boosting precision crop protection towards
agriculture 5.0 via machine learning and emerging technologies: A contextual review. Frontiers in Plant Science, 14.
doi:10.3389/fpls.2023.1143326.
[7] Kovács, I., & Husti, I. (2018). The role of digitalization in the agricultural 4.0 – how to connect the industry 4.0 to agriculture?
Hungarian Agricultural Engineering, 33(33), 38–42. doi:10.17676/hae.2018.33.38.
[8] Jha, K., Doshi, A., Patel, P., & Shah, M. (2019). A comprehensive review on automation in agriculture using artificial intelligence.
Artificial Intelligence in Agriculture, 2, 1–12. doi:10.1016/[Link].2019.05.004.
[9] Barkunan, S. R., Bhanumathi, V., & Sethuram, J. (2019). Smart sensor for automatic drip irrigation system for paddy cultivation.
Computers and Electrical Engineering, 73, 180–193. doi:10.1016/[Link].2018.11.013.
[10] Mottaleb, K. A. (2018). Perception and adoption of a new agricultural technology: Evidence from a developing country.
Technology in Society, 55, 126–135. doi:10.1016/[Link].2018.07.007.
[11] Sharma, R., Kamble, S. S., Gunasekaran, A., Kumar, V., & Kumar, A. (2020). A systematic literature review on machine learning
applications for sustainable agriculture supply chain performance. Computers and Operations Research, 119.
doi:10.1016/[Link].2020.104926.
[12] Kenny, E. M., Ruelle, E., Geoghegan, A., Shalloo, L., O’Leary, M., O’Donovan, M., Temraz, M., & Keane, M. T. (2021).
Bayesian Case-Exclusion and Explainable AI (XAI) for Sustainable Farming. The 29th International Joint Conference on
Artificial Intelligence - 17th Pacific Rim International Conference on Artificial Intelligence (IJCAI-PRICAI-20), 7-17 January,
2021, Yokohama, Japan.
[13] Schwab, K. (2017). The Global Competitiveness Report 2017-2018. World Economic Forum, Cologny, Switzerland.
[14] Minh, D., Wang, H. X., Li, Y. F., & Nguyen, T. N. (2022). Explainable artificial intelligence: a comprehensive review. Artificial
Intelligence Review, 55(5), 3503–3568. doi:10.1007/s10462-021-10088-y.
[15] Islam, S. R., Eberle, W., Ghafoor, S. K., & Ahmed, M. (2021). Explainable artificial intelligence approaches: A survey. arXiv
preprint arXiv:2101.09429. doi:10.48550/arXiv.2101.09429.
[16] Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D.,
Benjamins, R., Chatila, R., & Herrera, F. (2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities
and challenges toward responsible AI. Information Fusion, 58, 82–115. doi:10.1016/[Link].2019.12.012.
[17] Fernandez, A., Herrera, F., Cordon, O., Jose Del Jesus, M., & Marcelloni, F. (2019). Evolutionary fuzzy systems for explainable
artificial intelligence: Why, when, what for, and where to? IEEE Computational Intelligence Magazine, 14(1), 69–81.
doi:10.1109/MCI.2018.2881645.
[18] Samek, W., Wiegand, T., & Müller, K. R. (2017). Explainable artificial intelligence: Understanding, visualizing and interpreting
deep learning models. arXiv preprint arXiv:1708.08296. doi:10.48550/arXiv.1708.08296.
[19] Haefner, N., Wincent, J., Parida, V., & Gassmann, O. (2021). Artificial intelligence and innovation management: A review,
framework, and research agenda. Technological Forecasting and Social Change, 162. doi:10.1016/[Link].2020.120392.
[20] Montavon, G., Samek, W., & Müller, K. R. (2018). Methods for interpreting and understanding deep neural networks. Digital
Signal Processing: A Review Journal, 73, 1–15. doi:10.1016/[Link].2017.10.011.

Page | 756
 Emerging Science Journal | Vol. 8, No. 2

[21] Guidotti, R., Monreale, A., Ruggieri, S., Turini, F., Giannotti, F., & Pedreschi, D. (2018). A survey of methods for explaining
black box models. ACM Computing Surveys, 51(5). doi:10.1145/3236009.

[22] Adadi, A., & Berrada, M. (2018). Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI). IEEE
Access, 6, 52138–52160. doi:10.1109/ACCESS.2018.2870052.

[23] Ehsan, U., Wintersberger, P., Liao, Q. V., Mara, M., Streit, M., Wachter, S., Riener, A., & Riedl, M. O. (2021). Operationalizing
Human-Centered Perspectives in Explainable AI. Extended Abstracts of the 2021 CHI Conference on Human Factors in
Computing Systems. doi:10.1145/3411763.3441342.

[24] Rudin, C. (2019). Stop explaining black box machine learning models for high stakes decisions and use interpretable models
instead. Nature Machine Intelligence, 1(5), 206–215. doi:10.1038/s42256-019-0048-x.

[25] Tricco, A. C., Lillie, E., Zarin, W., O’Brien, K. K., Colquhoun, H., Levac, D., Moher, D., Peters, M. D. J., Horsley, T., Weeks,
… Straus, S. E. (2018). PRISMA extension for scoping reviews (PRISMA-ScR): Checklist and explanation. Annals of Internal
Medicine, 169(7), 467–473. doi:10.7326/M18-0850.
[26] Otieno, M. (2023). An extensive survey of smart agriculture technologies: Current security posture. World Journal of Advanced
Research and Reviews, 18(3), 1207–1231. doi:10.30574/wjarr.2023.18.3.1241.
[27] Race, D., Gentle, P., & Mathew, S. (2023). Living on the margins: Climate change impacts and adaptation by remote
communities living in the Pacific Islands, the Himalaya and desert Australia. Climate Risk Management, 40.
doi:10.1016/[Link].2023.100503.
[28] Rodríguez, J. P., Montoya-Munoz, A. I., Rodriguez-Pabon, C., Hoyos, J., & Corrales, J. C. (2021). IoT-Agro: A smart farming
system to Colombian coffee farms. Computers and Electronics in Agriculture, 190. doi:10.1016/[Link].2021.106442.
[29] Bwambale, E., Abagale, F. K., & Anornu, G. K. (2022). Smart irrigation monitoring and control strategies for improving water
use efficiency in precision agriculture: A review. Agricultural Water Management, 260. doi:10.1016/[Link].2021.107324.
[30] Manlove, J. L., Shew, A. M., & Obembe, O. S. (2021). Arkansas producers value upload speed more than download speed for
precision agriculture applications. Computers and Electronics in Agriculture, 190. doi:10.1016/[Link].2021.106432.
[31] Carrer, M. J., Filho, H. M. de S., Vinholis, M. de M. B., & Mozambani, C. I. (2022). Precision agriculture adoption and technical
efficiency: An analysis of sugarcane farms in Brazil. Technological Forecasting and Social Change, 177.
doi:10.1016/[Link].2022.121510.
[32] Thakur, N., Nigam, M., Mann, N. A., Gupta, S., Hussain, C. M., Shukla, S. K., Shah, A. A., Casini, R., Elansary, H. O., & Khan,
S. A. (2023). Host-mediated gene engineering and microbiome-based technology optimization for sustainable agriculture and
environment. Functional and Integrative Genomics, 23(1). doi:10.1007/s10142-023-00982-9.
[33] Said Mohamed, E., Belal, A. A., Kotb Abd-Elmabod, S., El-Shirbeny, M. A., Gad, A., & Zahran, M. B. (2021). Smart farming
for improving agricultural management. Egyptian Journal of Remote Sensing and Space Science, 24(3), 971–981.
doi:10.1016/[Link].2021.08.007.

[34] Adamides, G., Kalatzis, N., Stylianou, A., Marianos, N., Chatzipapadopoulos, F., Giannakopoulou, M., Papadavid, G., Vassiliou,
V., & Neocleous, D. (2020). Smart farming techniques for climate change adaptation in Cyprus. Atmosphere, 11(6).
doi:10.3390/ATMOS11060557.

[35] Shrivastava, A., Nayak, C. K., Dilip, R., Samal, S. R., Rout, S., & Ashfaque, S. M. (2023). Automatic robotic system design and
development for vertical hydroponic farming using IoT and big data analysis. Materials Today: Proceedings, 80, 3546–3553.
doi:10.1016/[Link].2021.07.294.

[36] Cordeiro, M., Markert, C., Araújo, S. S., Campos, N. G. S., Gondim, R. S., da Silva, T. L. C., & da Rocha, A. R. (2022). Towards
Smart Farming: Fog-enabled intelligent irrigation system using deep neural networks. Future Generation Computer Systems,
129, 115–124. doi:10.1016/[Link].2021.11.013.

[37] Fraser, A. (2022). ‘You can’t eat data’?: Moving beyond the misconfigured innovations of smart farming. Journal of Rural
Studies, 91, 200–207. doi:10.1016/[Link].2021.06.010.

[38] Nyangaresi, V. O., El-Omari, N. K. T., & Nyakina, J. N. (2022). Efficient Feature Selection and ML Algorithm for Accurate
Diagnostics. Journal of Computer Science Research, 4(1), 10–19. doi:10.30564/jcsr.v4i1.3852.

[39] Zambon, I., Cecchini, M., Egidi, G., Saporito, M. G., & Colantoni, A. (2019). Revolution 4.0: Industry vs. agriculture in a future
development for SMEs. Processes, 7(1). doi:10.3390/pr7010036.

[40] Polymeni, S., Plastras, S., Skoutas, D. N., Kormentzas, G., & Skianis, C. (2023). The Impact of 6G-IoT Technologies on the
Development of Agriculture 5.0: A Review. Electronics, 12(12), 2651. doi:10.3390/electronics12122651.

[41] Javaid, M., Haleem, A., Singh, R. P., & Suman, R. (2022). Enhancing smart farming through the applications of Agriculture 4.0
technologies. International Journal of Intelligent Networks, 3, 150–164. doi:10.1016/[Link].2022.09.004.

Page | 757
 Emerging Science Journal | Vol. 8, No. 2

[42] Naikwade, R. R., Patle, B. K., Joshi, V. S., Pagar, N. D., & Hirwe, S. B. (2021). Agriculture 5.0: Future of Smart Farming.
National Conference on Innovative Global Technology Trends in Art, Design, Technology, Management, Vedic Science,
Education and Architecture, Film & Media, 1-6.

[43] Mesías, F. J., Martín, A., & Hernández, A. (2021). Consumers’ growing appetite for natural foods: Perceptions towards the use
of natural preservatives in fresh fruit. Food Research International, 150. doi:10.1016/[Link].2021.110749.

[44] Zhang, A., Mankad, A., & Ariyawardana, A. (2020). Establishing confidence in food safety: is traceability a solution in
consumers’ eyes? Journal Fur Verbraucherschutz Und Lebensmittelsicherheit, 15(2), 99–107. doi:10.1007/s00003-020-01277-
y.

[45] Murugesan, R., Sudarsanam, S. K., Malathi, G., Vijayakumar, V., Neelanarayanan, V., Venugopal, R., Rekha, D., Saha, S., Bajaj,
R., Miral, A., & Malolan, V. (2019). Artificial intelligence and agriculture 5. 0. International Journal of Recent Technology and
Engineering, 8(2), 1870–1877. doi:10.35940/ijrte.B1510.078219.

[46] Ragazou, K., Garefalakis, A., Zafeiriou, E., & Passas, I. (2022). Agriculture 5.0: A New Strategic Management Mode for a Cut
Cost and an Energy Efficient Agriculture Sector. Energies, 15(9). doi:10.3390/en15093113.

[47] Berawi, M. A. (2019). Managing Nature 5.0 in industrial revolution 4.0 and society 5.0 era. International Journal of Technology,
10(2), 222–225. doi:10.14716/ijtech.v10i2.3084.

[48] Kwaghtyo, D. K., & Eke, C. I. (2023). Smart farming prediction models for precision agriculture: a comprehensive survey.
Artificial Intelligence Review, 56(6), 5729–5772. doi:10.1007/s10462-022-10266-6.

[49] Van Teeffelen, D. (2023). Plant disease detection with machine and deep learning: A systematic literature review & experimental
study. Master Thesis, Wageningen University and Research, Wageningen, Netherlands.

[50] Gardezi, M., Joshi, B., Rizzo, D. M., Ryan, M., Prutzer, E., Brugler, S., & Dadkhah, A. (2023). Artificial intelligence in farming:
Challenges and opportunities for building trust. Agronomy Journal, 1-12. doi:10.1002/agj2.21353.

[51] Vikranth, K., & Krishna Prasad, K. (2021). An Implementation of IoT and Data Analytics in Smart Agricultural System – A
Systematic Literature Review. International Journal of Management, Technology, and Social Sciences, 41–70.
doi:10.47992/ijmts.2581.6012.0129.

[52] Boursianis, A. D., Papadopoulou, M. S., Diamantoulakis, P., Liopa-Tsakalidi, A., Barouchas, P., Salahas, G., Karagiannidis, G.,
Wan, S., & Goudos, S. K. (2022). Internet of Things (IoT) and Agricultural Unmanned Aerial Vehicles (UAVs) in smart farming:
A comprehensive review. Internet of Things, 18, 100187. doi:10.1016/[Link].2020.100187.

[53] Dara, R., Hazrati Fard, S. M., & Kaur, J. (2022). Recommendations for ethical and responsible use of artificial intelligence in
digital agriculture. Frontiers in Artificial Intelligence, 5. doi:10.3389/frai.2022.884192.

[54] Bacco, M., Barsocchi, P., Ferro, E., Gotta, A., & Ruggeri, M. (2019). The Digitisation of Agriculture: a Survey of Research
Activities on Smart Farming. Array, 3–4, 100009. doi:10.1016/[Link].2019.100009.

[55] Klerkx, L., Jakku, E., & Labarthe, P. (2019). A review of social science on digital agriculture, smart farming and agriculture 4.0:
New contributions and a future research agenda. NJAS - Wageningen Journal of Life Sciences, 1-16.
doi:10.1016/[Link].2019.100315.

[56] Javed, A. R., Ahmed, W., Pandya, S., Maddikunta, P. K. R., Alazab, M., & Gadekallu, T. R. (2023). A Survey of Explainable
Artificial Intelligence for Smart Cities. Electronics (Switzerland), 12(4), 1020. doi:10.3390/electronics12041020.

[57] Yenduri, G., & Gadekallu, T. R. (2023). XAI for Maintainability Prediction of Software-Defined Networks. Proceedings of the
24th International Conference on Distributed Computing and Networking, 402-406. doi:10.1145/3571306.3571443.

[58] Du, M., Liu, N., & Hu, X. (2020). Techniques for interpretable machine learning. Communications of the ACM, 63(1), 68–77.
doi:10.1145/3359786.

[59] Antoniadi, A. M., Du, Y., Guendouz, Y., Wei, L., Mazo, C., Becker, B. A., & Mooney, C. (2021). Current challenges and future
opportunities for XAI in machine learning-based clinical decision support systems: A systematic review. Applied Sciences
(Switzerland), 11(11), 5088. doi:10.3390/app11115088.

[60] Slack, D., Hilgard, S., Jia, E., Singh, S., & Lakkaraju, H. (2020). Fooling LIME and SHAP. Proceedings of the AAAI/ACM
Conference on AI, Ethics, and Society, 180-186. doi:10.1145/3375627.3375830.

[61] Nauta, M., Trienes, J., Pathak, S., Nguyen, E., Peters, M., Schmitt, Y., Schlötterer, J., Van Keulen, M., & Seifert, C. (2023).
From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AI. ACM
Computing Surveys, 55(13), 1-42. doi:10.1145/3583558.

[62] Huuskonen, J., & Oksanen, T. (2018). Soil sampling with drones and augmented reality in precision agriculture. Computers and
Electronics in Agriculture, 154, 25–35. doi:10.1016/[Link].2018.08.039.

Page | 758
 Emerging Science Journal | Vol. 8, No. 2

[63] Atitallah, S. Ben, Driss, M., Boulila, W., & Ghezala, H. Ben. (2020). Leveraging Deep Learning and IoT big data analytics to
support the smart cities development: Review and future directions. Computer Science Review, 38.
doi:10.1016/[Link].2020.100303.

[64] van Etten, J., de Sousa, K., Cairns, J. E., Dell’Acqua, M., Fadda, C., Guereña, D., Heerwaarden, J. van, Assefa, T., Manners, R.,
Müller, A., Enrico Pè, M., Polar, V., Ramirez-Villegas, J., Øivind Solberg, S., Teeken, B., & Tufan, H. A. (2023). Data-driven
approaches can harness crop diversity to address heterogeneous needs for breeding products. Proceedings of the National
Academy of Sciences, 120(14), 1-10. doi:10.1073/pnas.2205771120.

[65] Paul, K., Chatterjee, S. S., Pai, P., Varshney, A., Juikar, S., Prasad, V., Bhadra, B., & Dasgupta, S. (2022). Viable smart sensors
and their application in data driven agriculture. Computers and Electronics in Agriculture, 198.
doi:10.1016/[Link].2022.107096.

[66] Ayoub Shaikh, T., Rasool, T., & Rasheed Lone, F. (2022). Towards leveraging the role of machine learning and artificial
intelligence in precision agriculture and smart farming. Computers and Electronics in Agriculture, 198.
doi:10.1016/[Link].2022.107119.

[67] Nasirahmadi, A., & Hensel, O. (2022). Toward the Next Generation of Digitalization in Agriculture Based on Digital Twin
Paradigm. Sensors, 22(2). doi:10.3390/s22020498.

[68] Chhetri, T. R., Hohenegger, A., Fensel, A., Kasali, M. A., & Adekunle, A. A. (2023). Towards improving prediction accuracy
and user-level explainability using deep learning and knowledge graphs: A study on cassava disease. Expert Systems with
Applications, 233. doi:10.1016/[Link].2023.120955.

[69] Hu, T., Zhang, X., Bohrer, G., Liu, Y., Zhou, Y., Martin, J., Li, Y., & Zhao, K. (2023). Crop yield prediction via explainable AI
and interpretable machine learning: Dangers of black box models for evaluating climate change impacts on crop yield.
Agricultural and Forest Meteorology, 336. doi:10.1016/[Link].2023.109458.

[70] Bandi, R., Swamy, S., & Arvind, C. S. (2023). Leaf disease severity classification with explainable artificial intelligence using
transformer networks. International Journal of Advanced Technology and Engineering Exploration, 10(100), 278–302.
doi:10.19101/IJATEE.2022.10100136.

[71] Chandra, H., Pawar, P. M., Elakkiya, R., Tamizharasan, P. S., Muthalagu, R., & Panthakkan, A. (2023). Explainable AI for Soil
Fertility Prediction. IEEE Access, 11, 97866–97878. doi:10.1109/ACCESS.2023.3311827.

[72] Sahidullah, M., Nayan, N. M., Morshed, M. S., Hossain, M. M., & Islam, M. U. (2023). Date Fruit Classification with Machine
Learning and Explainable Artificial Intelligence. International Journal of Computer Applications, 184(50), 1–5.
doi:10.5120/ijca2023922617.

[73] Celik, M. F., Isik, M. S., Taskin, G., Erten, E., & Camps-Valls, G. (2023). Explainable Artificial Intelligence for Cotton Yield
Prediction With Multisource Data. IEEE Geoscience and Remote Sensing Letters, 20. doi:10.1109/LGRS.2023.3303643.

[74] Bhat, S. A., Hussain, I., & Huang, N. F. (2023). Soil suitability classification for crop selection in precision agriculture using
GBRT-based hybrid DNN surrogate models. Ecological Informatics, 75. doi:10.1016/[Link].2023.102109.

[75] Ryo, M. (2022). Explainable artificial intelligence and interpretable machine learning for agricultural data analysis. Artificial
Intelligence in Agriculture, 6, 257–265. doi:10.1016/[Link].2022.11.003.

[76] Kawakura, S., Hirafuji, M., Ninomiya, S., & Shibasaki, R. (2022). Adaptations of Explainable Artificial Intelligence (XAI) to
Agricultural Data Models with ELI5, PDPbox, and Skater using Diverse Agricultural Worker Data. European Journal of Artificial
Intelligence and Machine Learning, 1(3), 27–34. doi:10.24018/ejai.2022.1.3.14.

[77] Cartolano, A., Cuzzocrea, A., Pilato, G., & Grasso, G. M. (2022). Explainable AI at Work! What Can It Do for Smart Agriculture?
2022 IEEE Eighth International Conference on Multimedia Big Data (BigMM). doi:10.1109/bigmm55396.2022.00020.

[78] Feldkamp, N., Genath, J., & Strassburger, S. (2022). Explainable AI For Data Farming Output Analysis: A Use Case for
Knowledge Generation Through Black-Box Classifiers. 2022 Winter Simulation Conference (WSC), Singapore.
doi:10.1109/wsc57314.2022.10015304.

[79] Mehedi, M. H. K., Hosain, A. K. M. S., Ahmed, S., Promita, S. T., Muna, R. K., Hasan, M., & Reza, M. T. (2022). Plant Leaf
Disease Detection using Transfer Learning and Explainable AI. 2022 IEEE 13th Annual Information Technology, Electronics
and Mobile Communication Conference (IEMCON), Vancouver, Canada. doi:10.1109/iemcon56893.2022.9946513.

[80] Kawakura, S., Hirafuji, M., Ninomiya, S., & Shibasaki, R. (2022). Analyses of Diverse Agricultural Worker Data with
Explainable Artificial Intelligence: XAI based on SHAP, LIME, and LightGBM. European Journal of Agriculture and Food
Sciences, 4(6), 11–19. doi:10.24018/ejfood.2022.4.6.348.

[81] Delaney, E. (2022). Case-based explanation for black-box time series and image models with applications in smart agriculture.
30th International Conference on Case-Based Reasoning, 11-12 September, 2022, Nancy, France.

Page | 759
 Emerging Science Journal | Vol. 8, No. 2

[82] Linardatos, P., Papastefanopoulos, V., & Kotsiantis, S. (2021). Explainable AI: A review of machine learning interpretability
methods. Entropy, 23(1), 1–45. doi:10.3390/e23010018.

[83] Zhang, Q., & Zhu, S. (2018). Visual interpretability for deep learning: a survey. Frontiers of Information Technology &
Electronic Engineering, 19(1), 27–39. doi:10.1631/fitee.1700808.

[84] Taj, I., & Zaman, N. (2022). Towards Industrial Revolution 5.0 and Explainable Artificial Intelligence: Challenges and
Opportunities. International Journal of Computing and Digital Systems, 12(1), 285–310. doi:10.12785/ijcds/120124.

[85] Selbst, A. D., Boyd, D., Friedler, S. A., Venkatasubramanian, S., & Vertesi, J. (2019). Fairness and Abstraction in Sociotechnical
Systems. Proceedings of the Conference on Fairness, Accountability, and Transparency, 56-68. doi:10.1145/3287560.3287598.

[86] Gonçalves, A. R., Pinto, D. C., Rita, P., & Pires, T. (2023). Artificial Intelligence and Its Ethical Implications for Marketing.
Emerging Science Journal, 7(2), 313-327. doi:10.28991/ESJ-2023-07-02-01.

[87] Saeed, W., & Omlin, C. (2023). Explainable AI (XAI): A systematic meta-survey of current challenges and future opportunities.
Knowledge-Based Systems, 263. doi:10.1016/[Link].2023.110273.

[88] Velten, S., Jager, N. W., & Newig, J. (2021). Success of collaboration for sustainable agriculture: a case study meta-analysis.
Environment, Development and Sustainability, 23(10), 14619–14641. doi:10.1007/s10668-021-01261-y.

[89] Maryono, M., Killoes, A. M., Adhikari, R., & Abdul Aziz, A. (2024). Agriculture development through multi-stakeholder
partnerships in developing countries: A systematic literature review. Agricultural Systems, 213. doi:10.1016/[Link].2023.103792.

[90] Hermans, F., Sartas, M., Van Schagen, B., Van Asten, P., & Schut, M. (2017). Social network analysis of multi-stakeholder
platforms in agricultural research for development: Opportunities and constraints for innovation and scaling. PLoS ONE, 12(2),
1-21. doi:10.1371/[Link].0169634.

[91] de Bruijn, H., Warnier, M., & Janssen, M. (2022). The perils and pitfalls of explainable AI: Strategies for explaining algorithmic
decision-making. Government Information Quarterly, 39(2), 101666. doi:10.1016/[Link].2021.101666.

[92] Gardezi, M., Adereti, D. T., Stock, R., & Ogunyiola, A. (2022). In pursuit of responsible innovation for precision agriculture
technologies. Journal of Responsible Innovation, 9(2), 224–247. doi:10.1080/23299460.2022.2071668.

[93] Rose, D. C., & Chilvers, J. (2018). Agriculture 4.0: Broadening Responsible Innovation in an Era of Smart Farming. Frontiers
in Sustainable Food Systems, 2. doi:10.3389/fsufs.2018.00087.

[94] Du, Y., Antoniadi, A. M., McNestry, C., McAuliffe, F. M., & Mooney, C. (2022). The Role of XAI in Advice-Taking from a
Clinical Decision Support System: A Comparative User Study of Feature Contribution-Based and Example-Based Explanations.
Applied Sciences, 12(20), 10323. doi:10.3390/app122010323.

[95] Jamil, H., Umer, T., Ceken, C., & Al-Turjman, F. (2021). Decision Based Model for Real-Time IoT Analysis Using Big Data
and Machine Learning. Wireless Personal Communications, 121(4), 2947–2959. doi:10.1007/s11277-021-08857-7.

[96] Weber, L., Lapuschkin, S., Binder, A., & Samek, W. (2023). Beyond explaining: Opportunities and challenges of XAI-based
model improvement. Information Fusion, 92, 154–176. doi:10.1016/[Link].2022.11.013.

Page | 760