The governance of artificial intelligence in agriculture: A review and future research directions

Abstract

Adopting artificial intelligence (AI) tools and technologies may be favourable for the agriculture sector in addressing food security and sustainability. As AI gains popularity, there are clear benefits to using AI in the agriculture sector. However, without effective governance, there are risks and challenges that can limit the benefits that can be obtained from using AI. For instance, ineffective policies that fail to outline liability in the event of adverse outcomes, such as crop loss resulting from ineffective decision-making by an AI technology or tool. We present a definition of AI governance in agriculture, in addition to a novel taxonomy which is comprised of nine principles: inclusivity & education, trustworthiness, custodianship & liability, law, sustainable & ethical development, transparency, privacy & security, bias, and data quality. Following a scoping review methodology, we conduct an analysis of the existing literature on AI governance in agriculture through the lens of the nine principles. Additionally, we uncover factors which support the principles of AI governance outlined in the discussion. Furthermore, identifying research gaps provides a clear road map for future directions, which can support researchers, policymakers, and practitioners. This study supports the ongoing development of AI governance and sheds light on the critical need for tailored governance for the agriculture sector.

Keywords

Artificial intelligence agriculture AI governance farmer-centric responsible innovation

Introduction

Emerging technologies, inclusive of artificial intelligence (AI), may be the solution to boosting the agriculture sector, supporting global food security and sustainability (Ahmad et al., 2024; Araújo et al., 2021; Green et al., 2021). The agriculture sector has become more information-rich, and as such, there has been a continued gradual shift towards the adoption of AI technologies (Khanna et al., 2024). AI, however, has been enlaced with technologies employed on farms for several years (Ryan, 2023; Wolfert et al., 2017), providing farmers with recommendations through data collection and analysis (Ryan, 2023).

AI continues to evolve alongside technologies such as drones and sensors, making its capabilities all the more significant (Ansari et al., 2024). To better understand what AI is, and how it can be applied in the agriculture sector, AI can be defined as ‘a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment’ (Organisation for Economic Co-operation and Development, 2019). By processing data collected from technologies, AI is able to make predictions, support in task automation, and optimize operations (Abuzar et al., 2024; Alaba et al., 2024; Ansari et al., 2024). With growing interest and research into the potential and applications of AI, studies have provided more focused insights into its impacts on the agricultural sector. AI can be utilized in various manners, such as to support plant disease detection (Rani et al., 2023), enhance real-time cattle identification (Mon et al., 2024), and optimize resource utilization in greenhouse operations (Mahdi et al., 2024). These advancements can support improved production and yield, allow for greater precision in input requirements, decrease wastage, enhance farmers’ well-being, and facilitate food safety and quality management, among others (Aijaz et a., 2025; Ryan et al., 2023; Sparrow et al., 2021; Taneja et al., 2023). The predictive capabilities of AI can be advantageous in overcoming shortcomings of existing data-driven models, as they enable farmers to predict outcomes of hypothetical scenarios they have not previously encountered (Neethirajan, 2023). Industry insights suggest that organizations within the agriculture sector are also showing an appetite for AI technologies, such as generative AI (Mayer et al., 2025; Shahriar et al., 2026). While the economic potential for organizations in agriculture may be smaller compared to other industries, such as healthcare, it is among the sectors with higher investment in generative AI (Mayer et al., 2025). Agriculture accounts for 4.1% of the global gross domestic product (World Bank, 2023), and employs roughly 26% of the world’s population (World Bank, 2025). With this growing demand, there is an opportunity for countries to develop their AI capacity. This has been observed in the United States, with Canada also considering potential approaches (Pasek et al., 2025).

While there is excitement surrounding the potential and benefits of AI, there remain challenges, concerns and risks. The rapid evolution of technologies, such as the application of AI, can be challenging for farmers as they need to keep up with infrastructure requirements, required financial capital, and skill requirements (Ansari et al., 2024; Ryan et al., 2023). Alongside these challenges, farmers face concerns about how AI will affect their privacy and security (Ansari et al., 2024). Concerns can also arise from a lack of transparency of AI tools and technologies, which can impact farmers’ trust in AI, potentially hindering its adoption (Dara et al., 2022). AI carries with it risks, which have been documented by Sparrow et al. (2021), Galaz et al. (2021), and Hampel and Fabulya (2024). The risks discussed include those commonly associated with technological advancements, such as job displacement, exploitation, surveillance, security, and alienation from the natural world. One might question what would happen if a company sells my data, or perhaps how I can ensure that the welfare of my livestock is not adversely impacted. To effectively address challenges, concerns and risks, there is a strong need for governance. AI governance needs to be rigorously developed into the design of the ecosystem of the technology to mitigate risks and eliminate challenges for farmers (Marchant and Allenby, 2017). It needs to be a comprehensive approach that addresses the entire AI lifecycle (problem statement, data collection and analysis, development, testing, deploying, and monitoring) (Cousineau et al., 2025).

Studies on AI governance can be seen across numerous sectors, including healthcare (Hassan et al., 2025; Schmidt et al., 2024; Stogiannos et al., 2023), agriculture (Alexander et al., 2024; Dara et al., 2022; Karanth et al., 2023), education (Filgueiras, 2024; Ghimire and Edwards, 2024), and more generally (Attard-Frost et al., 2024; Papagiannidis et al., 2025). This area of research is important as it allows for the critical review and understanding of what the challenges, concerns and risks are associated with AI, and what mechanisms can be put in place through governance to address them. Governance is a term with a number of definitions; however, it can generally be defined as ‘the setting and management of political rules of the game, and more substantially with a search for control, steering and accountability’ (Kjaer, 2023). It is an instrumental tool ensuring users can obtain the intended benefits of AI. AI governance can be achieved through tools, rules, processes, and other means, ensuring that the user or organization’s use, development, and delivery of AI tools and technologies aligns with certain principles (Birkstedt et al., 2023; Butcher and Beridze, 2019; Mäntymäki et al., 2022). These principles, addressing the challenges, concerns and risks, could include ensuring that AI is bias-free (Stogiannos et al., 2023), and that it aligns with privacy regulations (Dara et al., 2022).

Existing studies with a focus on the agriculture sector, such as those by Dara et al. (2022) and Alexander et al. (2024), use the term governance in discussion with AI. However, despite increased interest in AI governance in agriculture and ongoing research, there remain gaps in the literature pertaining to universal definitions and terminology, as well as a holistic review. This article aims to understand the current state of research on AI governance in the context of the agriculture sector. The discussion presented in this article supports ongoing efforts and developments by researchers, policymakers, and practitioners, not only by addressing research gaps but also by highlighting where limitations with current AI governance exist in the agriculture sector. Our contributions include providing a holistic review of AI governance in agriculture while aligning the review with the principles of AI governance. The review will:

Define AI governance in the context of agriculture and develop a taxonomy which guides the organization of the research captured in the scoping review.

Review the existing research which studies or discusses AI governance in agriculture.

Identify gaps in AI governance in agriculture that need to be addressed.

Propose future directions addressing gaps in existing research surrounding AI governance in agriculture.

This article is structured as follows: In the ‘Governance of AI in agriculture’ section, we review AI governance, define the term and provide readers with a taxonomy. An outline of the methodology employed in this scoping review is then provided in the ‘Methodology’ section. The ‘Results’ section provides an overview of the literature captured in the scoping review and presents the literature in alignment with the taxonomy. In the ‘Discussion’ section, we provide a discussion on various concepts of AI which are connected with AI governance, such as interdisciplinary development. Future directions based upon the research gaps are then provided in the ‘Research gaps and future directions’ section. Lastly, in the ‘Conclusion’ section, a conclusion is provided with a discussion on the contribution of this work, limitations and the importance of AI governance in agriculture.

Governance of AI in agriculture

The risks of AI have been well studied, and the existing work sheds light on how farmers and the agriculture sector could be impacted by it (Galaz et al., 2021; Hampel and Fabulya, 2024; Sparrow et al., 2021). Farmers can experience the risks of AI across daily data collection and decision-making processes, from AI applications and technologies (Galaz et al., 2021; Hampel and Fabulya, 2024; Sparrow et al., 2021). As farmers generate data from on-farm technologies, this information is often collected by the company that is providing the goods or services (Anidu and Dara, 2021; Dara et al., 2022). Without clear agreements detailing how data is collected, used, disseminated or retained and what data practices are implemented in the system, farmers are left vulnerable to exploitation. They risk the potential for companies to sell their data, which can later be used in ways that farmers did not know about or had not intended. Leveraging AI technology for data-driven decision-making applications can support farmers in determining input requirements for crop operations (Sparrow et al., 2021). However, if the output of the AI tool is incorrect, the farmer may have inaccurately managed its operation. This could result in crop loss and misalignment with regulatory compliance, among other outcomes. As such, this can have detrimental impacts on farmers, resulting in financial losses, for example. These are just two scenarios of risks associated with AI that can have a negative impact on farmers. To prevent these types of scenarios from occurring, there is a need to properly govern AI technologies. First, however, we must determine what AI governance means, what it encompasses, and how it differs in an agricultural context.

The term ‘AI governance’ generally lacks a single, universally agreed-upon definition (Attard-Frost et al., 2024), and this is true in an agriculture context as well. Instead, it can be defined in various manners and may be unique depending on its intended use. For instance, Papagiannidis et al. (2023) highlight research from Smuha (2019) and Amershi et al. (2019) that demonstrate how AI governance can be viewed from a technical software lens, as opposed to a human-centric and ethics-centric lens. AI governance can also be characterized as a set of tools, rules, processes, solutions, and other means that support the development, deployment, and usage of AI, aligning with certain principles and intentions (Birkstedt et al., 2023; Butcher and Beridze, 2019; Mäntymäki et al., 2022).

Here, we provide three definitions of AI governance:

AI governance is a system of rules, practices, processes, and technological tools that are employed to ensure an organization’s use of AI technologies aligns with the organization’s strategies, objectives, and values; fulfills legal requirements; and meets principles of ethical AI followed by the organization. (Mäntymäki et al., 2022)

A system of frameworks, practices and processes at an organizational level. AI governance helps various stakeholders implement, manage and oversee the use of AI technology. It also helps manage associated risks to ensure AI aligns with stakeholders’ objectives, is developed and used responsibly and ethically, and complies with applicable requirements. (IAPP Editorial Staff, 2025)

Effective AI governance requires guiding values and principles to ensure the trustworthy development and use of AI systems. (Organisation for Economic Co-operation and Development, 2025)

With an understanding of how AI governance is defined for other applications (not solely intended for AI governance in agriculture), we build on the existing definitions to provide one for the agriculture sector.

We define AI governance in the context of the agriculture and food sector as procedures, policies, standards, frameworks, technical solutions, ethical considerations, and methods that are required for the development, deployment, and maintenance of responsible AI systems.

We can further build on this definition with the goal that AI governance should support the development, deployment, and usage of AI in agriculture, ensuring sustainable and ethical usage and outcomes, minimizing potential risks and harm.

Having developed a definition for AI governance, we proceed to understanding what AI governance entails in agriculture, and what principles support effective AI governance. In doing so, we explore existing governance frameworks and models in the data and technology space, prior research, and resources through various organizations. When reviewing existing governance frameworks and models in the data and technology space, data governance is often mentioned. As highlighted by Ruder and Wittman (2025), data governance has been connected to AI in existing literature. Data governance integrates three elements: people, process, and technology, and has been examined in numerous use cases (Abbasi et al., 2016; Malik, 2013; Uren and Edwards, 2023). This includes research seeking to apply digital transformation in higher education (Taher, 2023), organizing data to support trustworthy AI (Janssen et al., 2020), evaluating the readiness of AI adoption (Bernardo et al., 2022), among other areas. Data governance has further looked more closely at aspects of data ownership, privacy, security, bias and data quality (Janssen et al., 2020; Rouzky et al., 2025; Ruder and Wittman, 2025). The elements are highly transferable, and more specifically, it is evident that within the context of technology, including AI, data is a critical component.

Prior research has identified that AI governance should consider transparency, algorithm bias, data privacy and protection, education, liability, ethical usage, ethical development, and the regulation of models (Mäntymäki et al., 2022; Ozor et al., 2025; Stogiannos et al., 2023). Additionally, resources and guidelines on the development of trustworthy and responsible AI can guide the understanding of AI governance. Organizations such as the National Institute of Standards and Technology (NIST) have released resources, providing clarity on what AI trustworthiness entails: validity and reliability, safety, security and resiliency, accountability and transparency, explainability and interpretability, privacy, and fairness with mitigation of harmful bias (National Institute of Standards and Technology, 2023; Theofanos et al., 2024). Resources from the NIST (National Institute of Standards and Technology, 2023; Theofanos et al., 2024), the Organisation for Economic Co-operation and Development (OECD) (Organisation for Economic Co-operation and Development, 2024), and the United Nations Educational, Scientific and Cultural Organization (UNESCO) (United Nations Educational, Scientific and Cultural Organization, n.d.), can support the development of AI technologies and tools (Ryan et al., 2025). These resources and guidelines allowed us to additionally consider aspects which contribute to trustworthy and responsible AI, and values-based principles for AI. All three of these organizations meaningfully contribute knowledge, resources, and soft law to the conversation on AI governance. As such, these resources were reviewed as part of our background research.

An understanding of what AI governance aims to achieve, address and resolve has allowed for the development of a taxonomy of AI governance which includes nine (9) principles: inclusivity & education, trustworthiness, custodianship & liability, law, sustainable & ethical development, transparency, privacy & security, bias, and data quality (Figure 1). Additionally, the principles are defined in Table 1. Having a definition for AI governance in agriculture, alongside the taxonomy, allows for a clear assessment of the current state of research on AI governance in the sector.

Figure 1.

A taxonomy of the nine (9) artificial intelligence (AI) governance principles.

Table 1.

Definitions of the nine principles of artificial intelligence (AI) governance.

Principle	Definition	Reference
Inclusivity & education	Individuals have an equal opportunity to access and use AI tools and technologies without experiencing barriers.	Ozor et al. (2025) and Bampasidou et al. (2024)
Trustworthiness	To use an AI tool or technology with confidence in its abilities, or to accept an output of the AI tool or technology without hesitation in its accuracy.	Birkstedt et al. (2023) and Karanth et al. (2023)
Custodianship &liability	An individual, organization or government that is responsible for AI governance being upheld and ensures that AI tools and technologies comply with AI governance throughout the AI lifecycle.	Birkstedt et al. (2023), Butcher and Beridze (2019),and Cousineau et al. (2025)
Law	Hard laws (legally binding) and soft laws (non-legally binding) can encompass, but are not limited to, laws, regulations, acts, policies, frameworks, and ethical considerations.	González-Rodríguez et al. (2024), Birkstedt et al. (2023), and Gwagwa et al. (2021)
Sustainable & ethicaldevelopment	Considerations are given to the environmental, social and ethical impact of AI tools and technologies during the AI lifecycle.	Stogiannos et al. (2023) and Camaréna (2021)
Transparency	AI tools and technologies that are clear and open throughout the AI lifecycle, specifically focused on how users’ interactions with AI tools and technologies will be owned and controlled by parties.	Cousineau et al. (2025) and Stogiannos et al. (2023)
Privacy & security	Measures that protect data from unauthorized access, cyberattacks, and data breaches, as well as safeguard personally identifiable information.	Dara et al. (2022) and Dhal and Kar (2025)
Bias	An output from an AI tool or technology that is discriminatory or stigmatizing as a result of flawed algorithm development, data availability, and data quality, among other factors, which can result in the output being unreliable.	Stogiannos et al. (2023), Cousineau et al. (2025), and Bellamy et al. (2019)
Data quality	Data used with and for AI tools and technologies is diverse, unbiased, credible, and of a high degree of accuracy.	Ozor et al. (2025) and Dhal and Kar (2025)

Methodology

Employing a scoping review method, the literature was identified using search terms relevant to the topic of AI governance in agriculture (Figure 2). The Web of Science Core Collection (https://www.webofscience.com/) was used to find the literature, which contained the following search terms in the titles (TI), abstracts (AB) or keywords (KW) (Ragany et al., 2023): artificial intelligence, AI, agricultur*, farm*, governanc*, trustworth*, responsib*, and ethic*.

Figure 2.

Indication of the number of studies reviewed and included in this work using the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) (Page et al., 2021). *Non-English records were in the following languages: Chinese, French, Latvian, and Spanish. Non-journal records were books, conferences, and series. A complete list of the included studies can be seen in Supplemental Table S1.

Sixteen combinations of search terms were used where artificial intelligence or AI was a required term in each search, along-side a secondary search term agricultur* or farm*, and a tertiary search term governanc*, trustworth*, responsib* or ethic*. The boolean search string was as follows: ((‘artificial intelligence’ OR ‘AI’) AND (agricultur* OR farm*) AND (governanc* OR trustworth* OR responsib* OR ethic*)). The search completed on 21 February 2025, yielded 281 records, of which 51 were removed to yield 230 publications for screening. Following further review for relevancy and search terms, 34 publications were selected for this work. It can be seen that there has been a gradual increase in research pertaining to AI governance in agriculture in recent years (Figure 3). Two investigators created the list of search terms and set the inclusion and exclusion criteria. One investigator completed the identification, screening, and inclusion steps (per PRISMA), which were then validated by a second investigator. Finally, two investigators developed the themes.

Figure 3.

The number of records from 2019 to 2025 included in this work. *Records were pulled 21 February 2025.

Results

In this section, we aim to review and understand the existing research regarding AI governance in agriculture, with the use of the nine AI governance principles: inclusivity & education, trustworthiness, custodianship & liability, law, sustainable & ethical development, transparency, privacy & security, bias, and data quality. While some of the literature may address more than one principle of AI governance, it may not appear on numerous occasions. This was done to highlight various works and create a more meaningful discussion. Additionally, there may be some principles that have been explored and addressed more frequently in existing research than others, which may provide a richer review and additional findings and perspectives.

Inclusivity & education

AI governance should ensure that AI technologies are inclusive, and resources in the form of education are accessible to stakeholders. When discussing AI governance in agriculture, it is important to identify the needs of the stakeholders to prepare for the changes associated with new technologies. Key risks which need to be addressed in the agriculture sector include workforce displacement (Bampasidou et al., 2024; Ryan, 2023) and the distribution of technologies (Bampasidou et al., 2024). The literature indicates that skilled agriculture workers are critical to supporting the development and effective use of AI technologies (Bampasidou et al., 2024; Holzinger et al., 2024; Tagkopoulos et al., 2024). Researchers discuss various methods for closing the skill gap by leveraging existing resources and creating new opportunities for enhancement and knowledge dissemination in the agricultural sector (Bampasidou et al., 2024; Holzinger et al., 2024; Tagkopoulos et al., 2024). While job displacement may be experienced, Bampasidou et al. (2024) and Ryan (2023) indicate that there may be a shift in what jobs require employees. In addition to the skill gap creating a barrier, research indicates that costs also pose a barrier. When considering costs, it must be acknowledged that costs go beyond the initial expense of the new AI technology (Bampasidou et al., 2024). Holzinger et al. (2024) further acknowledge that high costs associated with employing technologies can create a divide, impacting smaller farms. While the discussed barriers may be universal, geographic regions may experience them differently. The review of a case study presented by Gwagwa et al. (2021), focusing on African countries, highlighted barriers to inclusivity. In Africa, nearly half the workers in agriculture are women (Gwagwa et al., 2021). As such, Gwagwa et al. (2021) discuss gender discrimination as a barrier for women in African countries, which can limit access to knowledge, hindering technology adoption. Eliminating barriers and mitigating risks through effective AI governance is needed to ensure that AI technologies are inclusive and that education is provided to support the users.

Trustworthiness

The trustworthiness of AI technologies among stakeholders is critical to effective AI governance. As one aspect of trust in technology, Karanth et al. (2023) diligently points out that technology can make mistakes and can fail. With AI technologies still evolving and with certain ambiguities surrounding AI governance, stakeholders must be mindful of this potential risk. Risks in agriculture can include the potential for a tool to not achieve the required outcome, resulting in loss or injury (Tagkopoulos et al., 2022). When trust is lacking among stakeholders such as farmers, Karanth et al. (2023) highlight that it can indicate a more significant gap in the governance method, where legal parameters, for example, may not suitably address the concerns of its users. Karanth et al. (2023) further suggest that mistrust may also be a result of a lack of understanding of results provided by AI technologies, such as algorithms. Research by Chen et al. (2023) has focused on engaging in dialogue with farmers to gain a deeper understanding of how the perception of trustworthiness is associated with the adoption of technologies. Chen et al. (2023) determined that both a higher trust for the technology and a higher trust for data privacy protection can be linked to a higher likelihood of adoption of various technologies using AI. This finding demonstrates that trust among farmers is an important factor in AI adoption tendencies, alongside other factors such as cost. By leveraging methods which engage human input and capitalize on existing knowledge, trust can be built (Tagkopoulos et al., 2022). This human-centric approach can aid in ensuring that AI technologies are developed and deployed with trustworthiness in mind, allowing for meaningful AI technologies.

Custodianship & liability

Establishing custodianship and liability for AI technologies is critical to AI governance. AI technologies can pose various risks for farmers, for example, if AI tools were to cause injury to livestock or apply fertilizer and pesticides incorrectly, it could have negative financial impacts (Garske et al., 2021). When these types of risks are incurred by farmers, it is important to know who is liable. There is, however, a challenge because, as stated by Garske et al. (2021), there is a lack of clarity surrounding liability with AI in agriculture. As such, this section of the literature review examines the existing research surrounding custodianship and liability of AI in agriculture. It draws on discussions concerning the role of governments and understanding how existing research frames the responsibilities at the micro, meso and macro levels. In a multi-level approach, the macro-level is where the legal system lies, the meso-level is where organizations are involved, and the micro-level is where individuals and families are engaged (Ryan et al., 2025). Ryan et al. (2025) suggest that while challenges at these various levels have been identified, research strongly supports a multi-level approach to AI governance for agriculture. To manage the risks of AI, legislation is being established, such as the EU AI Act (Alexander et al., 2024). Currently, discussions by Alexander et al. (2024) highlight that research on AI technologies takes place under conditions that often lack ‘government oversight and regulation’, with the terms ‘ambiguity’ and ‘absence’ being used to describe the regulation and oversight. This demonstrates that there is a lack of custodianship at various points in the AI lifecycle. It is essential to consider this when establishing custodianship to support not only AI research but also the entire AI lifecycle, ensuring favourable outcomes and minimizing or mitigating risks. Clarity in liability when unfavourable outcomes are experienced as a result of using AI technologies in the agriculture sector is supported by effective custodianship.

Law

Law plays an instrumental role in AI governance. They allow for a form of standardization with which a user or organization should comply. Not all laws, however, are legally binding, and there could be a set of guidelines or best practices that a user or organization should consider. The distinction between legally binding (hard laws) and non-legally binding (soft laws) is used to review the existing discussion on this principle. It should also be noted that the law principle is intertwined with many of the other principles, such as trust. Without sufficient laws in place, challenges and risks can arise, such as a power imbalance between agriculture technology providers (ATPs) and farmers, which can place the farmer in a position where they have limited control or ownership over their data (Gardezi et al., 2024). Prior research has suggested that with the development of effective hard laws and soft laws, risks can be minimized, and trust can be built with farmers (Alexander et al., 2024; Gardezi et al., 2024). NIST and OECD have both released guidelines and recommendations for AI, specifically developing trustworthy and responsible AI (Alexander et al., 2024) and outlining five value-based principles of AI, respectively (Gardezi et al., 2024). The EU AI Act was the first of its kind to focus on mitigating and managing the risks of AI (Alexander et al., 2024). Other countries such as Canada and the United States have also put forward bills and acts (Gardezi et al., 2024). Gardezi et al. (2024) discuss that within the EU AI act, there is limited discussion on what risks the agriculture sector faces as a result of AI, as it is primarily focused on harm to humans.

Effective AI governance can have widespread impacts, such as interoperability and data sharing can be facilitated through common technology standards (González-Rodríguez et al., 2024), while regulations and legal agreements ensure privacy (Dara et al., 2022). A study by González-Rodríguez et al. (2024) calls for a mixed approach of hard laws and soft laws with a focus on AI phytopathology technologies in order to address current gaps, such as quality control and liability. There is also a need to ensure that governance and implementation are unified to limit barriers to research in the area of food systems (Alexander et al., 2024). It is also important to ensure that laws such AI ethics frameworks are developed with a multi-level approach in order to not overlook challenges that may be more complex (Ryan et al., 2025). As stated eloquently by Alexander et al. (2024) concerning the call and progress on ethical and regulatory frameworks, ‘these steps, while positive, indicate that the development of AI technologies has been rapidly moving forward within a vacuum of policy and consensus’. Gardezi et al. (2024) similarly indicate that AI innovations are evolving more rapidly than AI regulations. This stresses the need for timely AI governance, which is suitable for the agriculture sector.

Sustainable & ethical development

The sustainable and ethical development of AI is a crucial principle in AI governance. This combined approach to development ensures that AI technologies address the needs of stakeholders while also considering that they are developed with ethical, environmental, and social considerations in mind. This enables a mindful approach to AI development, allowing for the effective leveraging of its benefits in the agricultural sector. When implementing AI technologies in agriculture, it is essential to consider the potential risks associated with various equipment (Sood et al., 2022). There is, however, an opportunity to minimize the risks through the design of the equipment and testing (Sood et al., 2022). Sustainable and ethical development intersects with numerous aspects of AI governance, as demonstrated by the literature. As such, engagement with stakeholders is an important aspect of the sustainable and ethical development of AI technologies for agriculture. A stakeholder engagement approach allows for both the engagement of stakeholders in the development of AI technologies and the dissemination of knowledge of AI governance. There are various stakeholders in the agriculture sector, such as farmers, ATPs, and policymakers (Dara et al., 2022; Ryan, 2020). Stakeholder engagement between policymakers and farmers is an important dynamic considering the potential for dissemination of knowledge pertaining to AI technologies at large (Dara et al., 2022). There is also engagement with those developing AI technologies and tools, and a recent publication by van Hilten et al. (2025) presents a unique tool to support governance in AI, more specifically focusing on the ethical, legal, and social aspects (ELSA) of AI. The tool allows stakeholders, specifically those seeking to develop AI technologies, to see its alignment with ELSA (van Hilten et al., 2025). The two-tiered approach engages stakeholders through an intake survey and an interview, and provides stakeholders access to experts in this area of research and can support the development of technologies which align with AI governance (van Hilten et al., 2025).

When addressing sustainable and ethical development principles in the agriculture sector, the research explores the concept of sustainable food systems driven by AI technologies and tools, designed to enhance sustainability (Camaréna, 2021). Camaréna (2020) provides a comprehensive overview of existing design approaches that have been proposed by various research groups to approach AI technology design sustainably. Subsequent work by Camaréna (2021) discusses a bottom-up approach with co-design. Frameworks can include numerous sustainable design decisions, but there are limitations to various framework approaches and their ability to advise on various phases of the AI lifecycle (Mallinger and Baeza-Yates, 2024). Literature additionally explored sustainable design, but targeting a niche area of research, such as crop genomics (Wójcik-Gront et al., 2024). Data challenges are highlighted by Wójcik-Gront et al. (2024), such as data complexity. These challenges need to be addressed in order to achieve the benefits of sustainable agriculture in the area of crop genomics. Addressing sustainability with a focus on the environment, Sparrow et al. (2021) note that AI has environmental costs. These environmental costs can occur during the production of the necessary hardware and during AI powering (Sparrow et al., 2021). Building on recognition of the environmental costs of AI, Camaréna (2020) highlights the need for solutions to be evaluated as a whole system, with environmental and social costs, and the trade-offs made clear. Current developments demonstrate a push for sustainable and ethical development of AI; however, there remains an opportunity to explore this further in the agriculture sector.

Transparency

Transparency is a principle of AI governance that is interwoven with others and needs to be considered for people, processes, technology and data. An emerging topic in the literature in this review looked at transparency, particularly around the discussion of ownership and power. Power imbalance is a risk which can result from contractual agreements farmers have with ATPs for using various technologies (Gardezi et al., 2024). ATPs often retain ownership and control of the farm data which contributes to the power imbalance (Gardezi et al., 2024). Data can encompass information that is both non-identifiable and identifiable (Rahaman et al., 2024). With the manner in which ownership is currently approached in the agriculture sector, ATPs hold a great deal of power (Ryan, 2020). A study from Ryan (2020) discusses research on the power of agriculture big data analytics (ABDA). The unique position of ATPs allows for various types of power to be used, such as leadership power (Ryan, 2020). Through leadership power, ATPs are in a position to cause farmers to enter various agreements while overlooking informed consent and an outline of data ownership (Ryan, 2020). There are further challenges of data ownership for farmers, as even if the farmer has ownership of their data, it is possible that a royalty-free license may be retained by an ATP, granting them access to share farmers’ data (Uddin et al., 2024). The lack of transparency in various instances can be troubling and hinder the adoption of AI technologies. This is an area which requires clear governance methods, placing significance on being forthcoming about what AI technologies are collecting, how models are being trained, how information is being used, and who controls and owns the data, among other aspects.

Privacy & security

With the increasing need and generation of data from AI technologies, research has explored privacy and security. Improper security of data presents a significant risk, as it can contain not only data concerning the farming operation, but also personally identifiable information (Rahaman et al., 2024). Dara et al. (2022) define privacy as an individual’s right to control the use, share, and retain personally identifiable information. Security is defined as measures used to ‘prevent unauthorized access, data breaches, and cyberattacks’ (Dhal and Kar, 2025). It should be noted that the terminology used in this section, particularly the term privacy, is important as privacy is, in fact, protected by regulations and legal agreements (Dara et al., 2022). Papers discussing food safety and smart farming highlight privacy challenges such as the mismanagement of data and unauthorized access (Dhal and Kar, 2025; Rahaman et al., 2024). Additionally, security challenges can include secure data storage and data use consent (Dhal and Kar, 2025). It is also essential to understand the significance of privacy and security to each stakeholder. As discussed by Dara et al. (2022), it should be determined if stakeholders, such as farmers, have an understanding of what they are agreeing to with the use of various technologies, and how they can exercise certain aspects of privacy, to which they have a right, such as the control to pause data collection. To support AI in agriculture, researchers have issued a call to action for the need to establish laws, policies, and governance systems at various levels to address current challenges related to privacy and security (Gebresenbet et al., 2023; Uddin et al., 2024).

Bias & data quality

The principles of bias and data quality are often discussed together in the literature, and are presented together for that reason. With large amounts of data being generated and utilized in the agricultural sector, it is essential to address data quality and bias, with a focus on ensuring accuracy and minimizing potential biases. In research discussing AI’s application in food systems, a risk of biased training data can result in negative outcomes, such as inaccurate predictions (Dhal and Kar, 2025). The inaccurate predictions can impact decisions regarding compliance, suppliers, and more (Dhal and Kar, 2025). As AI technologies in agriculture can support decision-making and management decisions in various scopes, Uddin et al. (2024) highlight that it is critical to ensure the accuracy of the outcome or recommendation when using these technologies. Methods for achieving this start with understanding the training data for the AI model (Dhal and Kar, 2025). Dhal and Kar (2025) suggest that ensuring diverse sets of data is an initial strategy for improving accuracy while auditing AI systems, which is also needed for due diligence at certain times. Research by Tzachor et al. (2022) further emphasizes the importance of high-quality data, as biased or poor data can lead to ineffective technologies and erode trust among farmers. To address these concerns, Ozor et al. (2025) calls for ‘data quality assurance frameworks and standardized protocols’. While discussions in the context of agriculture may be limited, general methods from data and algorithms can be examined to determine if they meet the needs of the agricultural sector. By beginning to identify the needs for frameworks and guidelines, advancements will be well-positioned to address AI governance in agriculture.

This Results section provides an overview of the current state of AI governance principles in existing research on AI governance. It can be seen that some of the principles may have more extensive discussion than others. We do not infer that this demonstrates a lack of research on governing principles, but it demonstrates the interdisciplinary nature of AI governance, where principles often impact and influence one another. While research has steadily increased, there are further considerations to effective AI governance which should be considered to support the AI governance principles.

Discussion

Viewing AI governance in agriculture through a broad lens shows that research in this area has continued to increase steadily. However, there is ample opportunity to contribute further to this area of research, as seen by the small body of literature relevant to this study’s scope. To address the goal of understanding the current state of research on AI governance in the context of the agriculture sector, so far, a definition for AI governance in agriculture has been provided in this paper, alongside a taxonomy of the nine AI governance principles. Additionally, a review of the existing literature concerning AI governance in agriculture has been completed and forms the results section of this work. The findings from the results suggest that overall, AI governance in agriculture is in its initial stages. Given its nascent state, there are clear gaps in AI governance mechanisms, including laws, policies, regulations, and best practices. This creates greater exposure to risks already introduced, and stresses the importance of addressing the challenges and concerns of farmers.

A review of the existing literature highlighted various gaps in AI governance, which result in farmers being left vulnerable due to power imbalances. Power imbalances not only impact the individual farmer and operation, but the sector collectively. Power imbalances can be seen with farmers being put in a position with limited control or ownership of their data (Gardezi et al., 2024; Roussaki et al., 2023). It can also result in farmers entering agreements while overlooking informed consent and an outline of data ownership (Ryan, 2020). The power imbalances can also cause limitations to interoperability among systems and technologies (Roussaki et al., 2023). This can further impact data sharing, as research has shown pertaining to IoT and smart farming (Chhetri et al., 2024; Sullivan et al., 2024). With limitations, barriers and restrictions being placed on the farmer, primarily the data, it can hinder the development of the sector and result in exploitation of the farmer.

There is also a risk that farmers and other stakeholders employed in the agriculture sector will experience job displacement as a result of AI technologies and tools. While job displacement may be experienced, literature has touched on addressing the current skill gap, indicating a shift in what jobs require employees. With these risks, there is, however, the need to not only recognize the skill gap, but also labour availability, which varies depending on geographic area. For example, in Canada, reports indicate a skilled labour shortage (Green et al., 2021). A challenge or barrier to entry pertaining to the financial costs of AI technologies and tools was also identified, and should be considered when assessing the impacts of AI robotics, some of which could contribute to the displacement of jobs (van der Burg et al., 2024). While there are concerns surrounding these labour changes, we must also acknowledge that AI technologies, such as AI robots, may reduce labour costs by reducing labour requirements (Hernández et al., 2024).

Lastly, the literature review demonstrates the importance of building farmers’ trust in AI technologies and tools. The significance of trust lies in the fact that a lack of trust has been shown to hinder the adoption of AI technologies and tools among farmers (Chen et al., 2023; Dara et al., 2022). From the review, it is clear that trust is connected to several of the principles. Through effective AI governance, there is a need to improve the current lack of transparency surrounding AI technologies and tools, as well as improve data privacy (Chen et al., 2023; Dara et al., 2022). We must also ensure the usage of high-quality data for the development of AI tools and technologies (Tzachor et al., 2022). There is also the overall need to develop stronger AI governance through hard laws and soft laws, ensuring it aligns with the needs of the user, in this instance, the farmer (Alexander et al., 2024; Gardezi et al., 2024; Karanth et al., 2023). Improvements in these aspects of AI can support gaining the trust of farmers, thus improving the likelihood to adopt AI technologies and tools, and feel confident that they are obtaining the benefits of AI.

In order to develop effective AI governance for the agriculture sector, the significance and need for interdisciplinary and multi-level development and farmer involvement in design considerations will be discussed. A review of the results demonstrated that the principles of AI governance are highly interdisciplinary (Ryan et al., 2023). As such, the development of AI governance for the agriculture sector, as well as research that supports it, should use an interdisciplinary approach. This approach can involve engagement with a broad range of stakeholders in agriculture, bringing together researchers from various backgrounds, including social and natural science disciplines (Klerkx et al., 2019). This also allows researchers to create unified approaches to AI governance. We can capitalize on multiple perspectives when individuals with different backgrounds may attribute different meanings to concepts and terminology (Prutzer et al., 2023). An interdisciplinary approach can also yield findings that address the needs of both the public and private sectors (Ryan et al., 2025). Research often discusses engagement with government and regulatory agencies, agriculture technology providers (at any stage of the AI lifecycle (Cousineau et al., 2025)), farmers, producers, growers, and research institutes. While each stakeholder is critical to effective AI governance, we suggest that farmer associations and agriculture NGOs also be considered during the development, deployment and usage of AI. Farmer associations and agriculture NGOs are often present in the agriculture sector at various levels (national, regional, local, etc.). Given the existing relationship with farmers, they can support and contribute to AI governance through the development of guidelines, programming, services, and courses, among others. This can further contribute to creating an interdisciplinary environment for developing effective AI governance methods for the agriculture sector. Research by Sigfrids et al. (2023) highlights the assumption that engagement from all stakeholders can lead to improvements in AI governance, whether through increased trust or more effective policy development.

Trustworthiness is a critical aspect not only of the adoption of AI in agriculture but also of effective AI governance. It can be built by meaningfully engaging stakeholders and leveraging their learned knowledge (Tagkopoulos et al., 2022). In particular, there should be strong farmer engagement, leveraging a farmer-centric approach. Drawing on research from (McCaig et al., 2023), a farmer-centric design approach encompasses discussions from the literature review and contributes effectively to the discussion on AI governance. Farmer-centric design thinking addresses four needs: technology solution, human in the loop, process and protocol, and policy and legal requirements. Effectively, this approach would allow AI providers to ensure that AI tools and technologies align with the needs and demands of farmers, while boosting trust as farmers are engaged from the start. This approach allows for various considerations, such as the robotic and human interactions, privacy preservation and explainability. van der Burg et al. (2024) found that respondents reflected on the ‘value of the evolving robot-human collaboration’, focusing on what facilitates appropriate interactions rather than the intellectual potentials of robot versus human. Additionally, with producers expressing concerns about confidentiality and data privacy (Chen et al., 2023; Dara et al., 2022), there is a need to explore privacy-preserving techniques further. Research has also found that explainability is important for AI models, as increased trust among farmers is achieved when they are presented in an appropriate manner (Dara et al., 2022; Rai, 2020). However, there remains room to improve upon how AI models’ decision-making process is presented (Dhal and Kar, 2025). As greater engagement is achieved with the stakeholders and greater emphasis is placed on leveraging a farmer-centric design approach for AI governance, there is opportunity to achieve harmony among the nine outline principles of AI governance, allowing all stakeholders to benefit from its potential.

Research gaps and future directions

Following the scoping review, research gaps and future directions were identified. This was done through the review of existing literature presented in the results and raised in the discussion. When addressing the research gaps and future directions, consideration should be given to the interdisciplinary nature of AI governance and stakeholder engagement, as they can positively impact AI governance overall.

The social implications of AI technology adoption: The first research gap identified is regarding the adoption of AI technologies and tools. This was identified primarily through the principles of inclusivity & education, and trustworthiness, where existing research did identify the risks of disparity among farmers due to their potential or ability to adopt AI technologies and tools. Trust among farmers in AI technologies and data privacy protection can create hesitancy to adopt AI (Chen et al., 2023; Dara et al., 2022). There is a need to gain a deeper understanding of the current information which is accessible to farmers with various AI technologies and to assess the current methods used by agriculture technology providers to be forthcoming with information concerning their AI technologies. The relationship between both stakeholders requires strong governance to improve upon existing relationships and create a positive outcome. A governance method that supports the standardization of aspects of data collection, usage, ownership, model training, technology agreement transparency, and so forth should be explored. In doing so, there is an opportunity to mitigate various risks, including the potential for disparities to emerge among farming communities.

Effective frameworks of AI governance: The second research gap we have identified is in regard to AI governance measures, particularly pertaining to the lack of frameworks. This gap primarily became apparent when reviewing the literature concerning the law principle, with a lack of existing hard laws and soft laws focused on the agriculture sector. This research has discussed some of the existing acts and bills put forward by countries surrounding AI technologies and tools (Alexander et al., 2024; Gardezi et al., 2024). There are limitations to some of the existing acts and bills, as they focus primarily on harm to humans, not the agriculture sector in particular. While there is a desire for farmers to experience the benefits of AI, there is a need to address the ambiguity or gap in the existing acts and bills in order to address the concerns, challenges and risks faced by farmers when using AI technologies and tools. However, it is unclear how much AI governance should address, where a balance is maintained between AI technologies and tools being effective and efficient, while addressing the concerns, challenges and risks. There is an opportunity to conduct a more comprehensive review of existing policies and acts for AI technologies and tools, and determine where there remain gaps in addressing what farmers deem risks and challenges. This should be assessed alongside how the policies and acts impact the effectiveness and efficiency of AI technologies and tools. This can allow for a clearer understanding of where trade-offs exist currently, and what trade-offs farmers may or may not be willing to come up against.

Developing suitable best practices: The third research gap we have identified is in regard to best practices for AI governance in the agriculture sector. This gap was also identified through the literature concerning the law principle. Best practices are an effective tool which can support the development, deployment and usage of AI technologies and tools. The literature review touched on organizations that have released guidelines and recommendations for trustworthy and responsible AI and discussed various risks of AI in agriculture (Alexander et al., 2024; Gardezi et al., 2024; Ryan et al., 2025). From the literature and existing resources, there remains an opportunity to capitalize on the stakeholder dynamics and develop best practices for AI governance in agriculture. Future research should explore the potential for farmer associations and agriculture NGOs to contribute to the development of best practices, similar to how the EU Code of conduct on agricultural data sharing by contractual agreement (Cogeca, 2018; Ryan et al., 2024) was developed with stakeholder engagement. Farmer associations and agriculture NGOs are in a unique position to disseminate information and influence farmers’ decision-making regarding the technologies they use on their operations.

Operationalization of AI governance: The fourth research gap focuses on the development of AI technologies and tools, while leveraging tools that support effective AI governance. Here, through the review of the sustainable & ethical development principle and the law principle, we identified a need to understand how tools and frameworks can aid developers, and how existing tools and frameworks address AI governance. Advancements in AI governance have enabled the creation of various tools for stakeholders, mainly developers, to assess the governance of their technologies (van Hilten et al., 2025). It is unclear from existing research how effectively existing AI technologies and tools align with AI governance and the emerging acts and policies. As existing AI governance may not be specific to the agriculture sector, it may allow for the identification of where either current AI governance methods need to be expanded, or new AI governance needs to be developed and used in conjunction with existing methods. This opens the opportunity to leverage tools, namely for developers of AI technologies and tools. A unique tool developed to assess the ELSA of AI has been previously discussed (van Hilten et al., 2025) and is already targeted at the agriculture sector. While developers are not required to use tools such as ELSA, future research should focus on understanding how these tools support ongoing AI governance efforts. It should aim to understand how these tools align with AI acts and regulations, and can possibly simplify developers’ efforts in assessing whether the AI technologies and tools developed are compliant. Further consideration can be given to developing a certificate or label that indicates to farmers that the AI technology or tool was developed under AI governance, thereby supporting their selection of what to adopt.

Farmer-centric technology and process solutions: The final research gap is regarding the current lack of research on the development of AI technologies and tools following a farmer-centric design approach. Farmers, as the users of the AI technologies and tools, should be actively part of the AI lifecycle, as highlighted in the results and discussion. A farmer-centric approach to AI tool and technology development shows promise in aligning with AI governance. The approach allows for technologies to be built and developed with the farmers in mind (McCaig et al., 2023). Applied to the agriculture sector, it allows for consideration to be given to: the user’s perception of technology, factors that need to be established to support adoption of said technology, the usability and reception from farmers of the technology, and lastly, the governance of the technology (McCaig et al., 2023). Future research should consider a living lab approach, which may be suitable for developing farmer-centric technologies, as it would allow for the development and testing of various solutions while addressing AI governance. Living labs have been extensively examined by Berberi et al. (2026) to understand what enables a well-functioning living lab, versus what creates potential barriers (Berberi et al., 2026). As solutions are tested, there is an opportunity to see how they align with the needs of farmers, and they can also support addressing various risks of AI technologies identified by existing research. The involvement of various stakeholders can also encourage an interdisciplinary approach, which, as research has indicated, is a beneficial approach for AI governance.

Conclusion

As AI technologies continue to evolve, growth will be seen among its users and in its applications. The agriculture sector has the potential to benefit from the evolving and iterative nature of AI technologies (Tzachor et al., 2022). Seeing as how there is the potential for new risks to emerge over time, and the current trend of research indicating its timely nature, we believe this study can contribute to ongoing research efforts. In this literature review, we aimed to provide clarity surrounding what AI governance in agriculture means, review the existing research, identify gaps, and outline future directions. Building on existing definitions of AI governance, the following definition is developed for the agriculture sector: AI governance in the context of the agriculture and food sector as procedures, policies, standards, frameworks, technical solutions, ethical considerations, and methods that are required for the development, deployment, and maintenance of responsible AI systems. Additionally, we develop a novel taxonomy of the nine principles of AI governance: inclusivity & education, trustworthiness, custodianship & liability, law, sustainable & ethical development, transparency, privacy & security, bias, and data quality. Through the review of the literature, we also find that there are factors which can support the nine principles for effective AI governance, primarily through using an interdisciplinary approach with involvement from all stakeholders. Our review of the literature reveals gaps in the existing literature, such as the social implications of AI technology adoption, developing suitable best practices and farmer-centric technology and process solutions. The identified gaps, if left unanswered, could hinder researchers, policymakers, and practitioners’ ability to propose effective AI governance measures in a timely manner, specifically tailored to the agriculture sector. While this study provides a foundation for future researchers and stakeholders to explore AI governance in agriculture, there are some limitations to the study. Despite a gradual increase in research about our topic, only a small body of literature contributed to this study. This limited the ability to use topic modelling (Wheeler et al., 2020). Additionally, there is an opportunity to expand on the literature included, in order to include grey literature and policy reports. We also acknowledge that between the time of completing the literature search and pull, and the time of publication, there may be relevant literature that is published, but not included in the scope of this work. With AI’s potential to, among other benefits, improve the precision of input requirements, enhance farmers’ well-being, and improve quality management (Aijaz et a., 2025; Ryan et al., 2023; Sparrow et al., 2021; Taneja et al., 2023), the future of the agriculture sector boasts many opportunities. We must, however, ensure that AI governance is at the forefront of AI technologies and tools, to ensure concerns, challenges and risks faced by farmers are mitigated or removed.

Supplemental Material

sj-pdf-1-oag-10.1177_00307270261450225 - Supplemental material for The governance of artificial intelligence in agriculture: A review and future research directions

Supplemental material, sj-pdf-1-oag-10.1177_00307270261450225 for The governance of artificial intelligence in agriculture: A review and future research directions by Michelle Ragany, Sjaak Wolfert and Rozita Dara in Outlook on Agriculture

Footnotes

Acknowledgements

Not applicable.

ORCID iDs

Michelle Ragany

Sjaak Wolfert

Rozita Dara

Ethical approval and informed consent statement

Not applicable.

Consent to participate

Not applicable.

Consent to publication

Not applicable.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant and Ontario Ministry of Agriculture Food and Rural Affairs funding awarded to Dr Rozita Dara.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

Not applicable.

Supplemental material

Supplemental material for this article is available online.

References

Abbasi

Sarker

Chiang

(2016) Big data research in information systems: Toward an inclusive research agenda. Journal of the Association for Information Systems 17(2): 3.

Abuzar

Kumar

Gupta

(2024) Emerging trends in IoT, AI and agricultural science. In: Agriculture 4.0. London: CRC Press, 373–390.

Ahmad

Liew

Venturini

, et al. (2024) AI can empower agriculture for global food security: Challenges and prospects in developing nations. Frontiers in Artificial Intelligence 7: 1328530.

Aijaz

Lan

Raza

, et al. (2025) Artificial intelligence in agriculture: Advancing crop productivity and sustainability. Journal of Agriculture and Food Research 20: 101762.

Alaba

Jegede

Sani

, et al. (2024) Artificial intelligence of things (AIoT) solutions for sustainable agriculture and food security. In: Artificial Intelligence of Things for Achieving Sustainable Development Goals. Cham: Springer, 123–142.

Alexander

Yarborough

Smith

(2024) Who is responsible for ‘responsible AI’?: Navigating challenges to build trust in AI agriculture and food system technology. Precision Agriculture 25(1): 146–185.

Amershi

Begel

Bird

, et al. (2019) Software engineering for machine learning: A case study. In: 2019 IEEE/ACM 41st international conference on software engineering: Software engineering in practice (ICSE-SEIP), Montreal, QC, Canada. Piscataway, NJ: IEEE, 291–300.

Anidu

Dara

(2021) A review of data governance challenges in smart farming and potential solutions. In: 2021 IEEE international symposium on technology and society (ISTAS), Waterloo, ON, Canada. Piscataway, NJ: IEEE, 1–8.

Ansari

Gupta

Kumar

, et al. (2024) Adoption challenges and incentives for IoT and AI in agriculture. In: Agriculture 4.0. London: CRC Press, 355–372.

10.

Araújo

Peres

Barata

, et al. (2021) Characterising the agriculture 4.0 landscape—Emerging trends, challenges and opportunities. Agronomy 11(4): 667.

11.

Attard-Frost

Brandusescu

Lyons

(2024) The governance of artificial intelligence in Canada: Findings and opportunities from a review of 84 AI governance initiatives. Government Information Quarterly 41(2): 101929.

12.

Bampasidou

Goldgaber

Gentimis

, et al. (2024) Overcoming ‘digital divides’: Leveraging higher education to develop next generation digital agriculture professionals. Computers and Electronics in Agriculture 224: 109181.

13.

Bellamy

Dey

Hind

, et al. (2019) AI fairness 360: An extensible toolkit for detecting and mitigating algorithmic bias. IBM Journal of Research and Development 63(4): 4/5–1.

14.

Berberi

Beaudoin

McPhee

, et al. (2026) Enablers, barriers, and future considerations for living lab effectiveness in environmental and agricultural sustainability transitions: A review of studies evaluating living labs. Local Environment 31(1): 1–19.

15.

Bernardo

Barroso

Santos

(2022) Artificial intelligence and digital forensics on data governance breaking through its importance to organizations and its operations. Forensic Science & Addiction Research 5(4): 437–39.

16.

Birkstedt

Minkkinen

Tandon

, et al. (2023) AI governance: Themes, knowledge gaps and future agendas. Internet Research 33(7): 133–167.

17.

Butcher

Beridze

(2019) What is the state of artificial intelligence governance globally? The RUSI Journal 164(5-6): 88–96.

18.

Camaréna

(2020) Artificial intelligence in the design of the transitions to sustainable food systems. Journal of Cleaner Production 271: 122574.

19.

Camaréna

(2021) Engaging with artificial intelligence (AI) with a bottom-up approach for the purpose of sustainability: Victorian Farmers Market Association, Melbourne, Australia. Sustainability 13(16): 9314.

20.

Chen

Cate

Cheren

(2023) Communicating agriculture AI technologies: How American agricultural producers’ perception of trustworthiness, risk perception, and emotion affect their likelihood of adopting artificial intelligence in food systems. Environmental Communication 17(8): 1004–1019.

21.

Chhetri

Dehury

Varghese

, et al. (2024) Enabling privacy-aware interoperable and quality IoT data sharing with context. Future Generation Computer Systems 157: 164–179.

22.

Cogeca

, et al. (2018) EU code of conduct on agricultural data sharing by contractual agreement.

23.

Cousineau

Dara

Chowdhury

(2025) Trustworthy AI: AI developers’ lens to implementation challenges and opportunities. Data and Information Management 9(2): 100082.

24.

Dara

Hazrati Fard

Kaur

(2022) Recommendations for ethical and responsible use of artificial intelligence in digital agriculture. Frontiers in Artificial Intelligence 5: 884192.

25.

Dhal

Kar

(2025) Leveraging artificial intelligence and advanced food processing techniques for enhanced food safety, quality, and security: A comprehensive review. Discover Applied Sciences 7(1): 75.

26.

Filgueiras

(2024) Artificial intelligence and education governance. Education, Citizenship and Social Justice 19(3): 349–361.

27.

Galaz

Centeno

Callahan

, et al. (2021) Artificial intelligence, systemic risks, and sustainability. Technology in Society 67: 101741.

28.

Gardezi

Joshi

Rizzo

, et al. (2024) Artificial intelligence in farming: Challenges and opportunities for building trust. Agronomy Journal 116(3): 1217–1228.

29.

Garske

Bau

Ekardt

(2021) Digitalization and AI in European agriculture: A strategy for achieving climate and biodiversity targets? Sustainability 13(9): 4652.

30.

Gebresenbet

Bosona

Patterson

, et al. (2023) A concept for application of integrated digital technologies to enhance future smart agricultural systems. Smart Agricultural Technology 5: 100255.

31.

Ghimire

Edwards

(2024) From guidelines to governance: A study of AI policies in education. In: International conference on artificial intelligence in education, Recife, Brazil. Cham: Springer, 299–307.

32.

González-Rodríguez

Izquierdo-Bueno

Cantoral

, et al. (2024) Artificial intelligence: A promising tool for application in phytopathology. Horticulturae 10(3): 197.

33.

Green

Abdulai

Duncan

, et al. (2021) A scoping review of the digital agricultural revolution and ecosystem services: Implications for Canadian policy and research agendas. Facets 6(1): 1955–1985.

34.

Gwagwa

Kazim

Kachidza

, et al. (2021) Road map for research on responsible artificial intelligence for development (AI4D) in African countries: The case study of agriculture. Patterns 2(12): 100381.

35.

Hampel

Fabulya

(2024) The risks of AI in agriculture. Analecta Technica Szegedinensia 18(4): 32–44.

36.

Hassan

Borycki

Kushniruk

(2025) Artificial intelligence governance framework for healthcare. In: Healthcare Management Forum. Vol. 38. Los Angeles, CA: SAGE Publications Sage CA, 125–130.

37.

Hernández

VEG

Pérez-Mesa

de Burgos Jiménez

, et al. (2024) Smart agriculture as source of sustainability in intensive greenhouse production. International Food and Agribusiness Management Review 27(5): 882–902.

38.

Holzinger

Fister

Kaul

, et al. (2024) Human-centered AI in smart farming: Toward agriculture 5.0. IEEE Access 12: 62199–62214.

39.

IAPP Editorial Staff (2025) Key terms for AI governance. Available at: https://iapp.org/resources/article/key-terms-for-ai-governance/ (updated version of October 2023 glossary).

40.

Janssen

Brous

Estevez

, et al. (2020) Data governance: Organizing data for trustworthy artificial intelligence. Government Information Quarterly 37(3): 101493.

41.

Karanth

Benefo

Patra

, et al. (2023) Importance of artificial intelligence in evaluating climate change and food safety risk. Journal of Agriculture and Food Research 11: 100485.

42.

Khanna

Atallah

Heckelei

, et al. (2024) Economics of the adoption of artificial intelligence-based digital technologies in agriculture. Annual Review of Resource Economics 16: 41–61.

43.

Kjaer

(2023) Governance. Cambridge: Polity.

44.

Klerkx

Jakku

Labarthe

(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 90: 100315.

45.

Mahdi

Mahmood

Tang

, et al. (2024) Prediction of energy efficiency and environmental sustainability in greenhouse tomato production: Insights and analysis. In: 2024 IEEE international conference on cognitive computing and complex data (ICCD), Qinzhou, China. Piscataway, NJ: IEEE, 424–428.

46.

Malik

(2013) Governing big data: Principles and practices. IBM Journal of Research and Development 57(3/4): 1–1.

47.

Mallinger

Baeza-Yates

(2024) Responsible AI in farming: A multi-criteria framework for sustainable technology design. Applied Sciences 14(1): 437.

48.

Mäntymäki

Minkkinen

Birkstedt

, et al. (2022) Defining organizational AI governance. AI and Ethics 2(4): 603–609.

49.

Marchant

Allenby

(2017) Soft law: New tools for governing emerging technologies. Bulletin of the Atomic Scientists 73(2): 108–114.

50.

Mayer

Yee

Chui

, et al. (2025) Superagency in the workplace. McKinsey & Company report. Empowering people to unlock AI’s full potential.

51.

McCaig

Dara

Rezania

(2023) Farmer-centric design thinking principles for smart farming technologies. Internet of Things 23: 100898.

52.

Mon

Onizuka

Tin

, et al. (2024) AI-enhanced real-time cattle identification system through tracking across various environments. Scientific Reports 14(1): 17779.

53.

National Institute of Standards and Technology (2023) Trustworthy and responsible AI. Available at: https://www.nist.gov/trustworthy-and-responsible-ai.

54.

Neethirajan

(2023) The significance and ethics of digital livestock farming. AgriEngineering 5(1): 488–505.

55.

Organisation for Economic Co-operation and Development (2019) Recommendation of the council on artificial intelligence. Technical Report OECD/LEGAL/0449. Organisation for Economic Co-operation and Development. https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449 (adopted 22 May 2019; amended 3 May 2024).

56.

Organisation for Economic Co-operation and Development (2024) OECD AI principles. Available at: https://www.oecd.org/en/topics/sub-issues/ai-principles.html.

57.

Organisation for Economic Co-operation and Development (2025) Steering AI’s future: Strategies for anticipatory governance. OECD artificial intelligence papers, no. 32. Organisation for Economic Co-operation and Development. Available at: https://www.oecd.org/publications/steering-ai-s-future-5480ff0a-en.htm.

58.

Ozor

Nwakaire

Nyambane

, et al. (2025) Enhancing Africa’s agriculture and food systems through responsible and gender inclusive AI innovation: Insights from AI4AFS network. Frontiers in Artificial Intelligence 7: 1472236.

59.

Page

McKenzie

Bossuyt

, et al. (2021) The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 372: n71.

60.

Papagiannidis

Enholm

Dremel

, et al. (2023) Toward AI governance: Identifying best practices and potential barriers and outcomes. Information Systems Frontiers 25(1): 123–141.

61.

Papagiannidis

Mikalef

Conboy

(2025) Responsible artificial intelligence governance: A review and research framework. The Journal of Strategic Information Systems 34(2): 101885.

62.

Pasek

Bronson

Centivany

, et al. (2025) Challenges and opportunities for a made-in-Canada approach to artificial intelligence. White paper, University of Ottawa.

63.

Prutzer

Gardezi

Rizzo

, et al. (2023) Rethinking ‘responsibility’ in precision agriculture innovation: Lessons from an interdisciplinary research team. Journal of Responsible Innovation 10(1): 2202093.

64.

Ragany

Haggag

El-Dakhakhni

, et al. (2023) Closed-loop agriculture systems meta-research using text mining. Frontiers in Sustainable Food Systems 7: 1074419.

65.

Rahaman

Lin

Pappachan

, et al. (2024) Privacy-centric AI and IoT solutions for smart rural farm monitoring and control. Sensors 24(13): 4157.

66.

Rai

(2020) Explainable AI: From black box to glass box. Journal of the academy of marketing science 48(1): 137–141.

67.

Rani

Sahoo

Bellamkonda

, et al. (2023) Role of artificial intelligence in agriculture: An analysis and advancements with focus on plant diseases. IEEE Access 11: 137999.

68.

Roussaki

Doolin

Skarmeta

, et al. (2023) Building an interoperable space for smart agriculture. Digital Communications and Networks 9(1): 183–193.

69.

Rouzky

Sarrafzadeh

Sowells-Boone

, et al. (2025) Data governance in smart farming: A systematic review of challenges and gaps. Information Processing in Agriculture.

70.

Ruder

Wittman

(2025) Agricultural data governance from the ground up: Exploring data justice with agri-food movements. Big Data & Society 12(1): 20539517251330182.

71.

Ryan

(2020) Agricultural big data analytics and the ethics of power. Journal of Agricultural and Environmental Ethics 33(1): 49–69.

72.

Ryan

(2023) The social and ethical impacts of artificial intelligence in agriculture: Mapping the agricultural AI literature. AI & Society 38(6): 2473–2485.

73.

Ryan

Atik

Rijswijk

, et al. (2024) The future of agricultural data-sharing policy in Europe: Stakeholder insights on the EU code of conduct. Humanities and Social Sciences Communications 11(1): 1–15.

74.

Ryan

De Roo

Wang

, et al. (2025) AI through the looking glass: An empirical study of structural social and ethical challenges in AI. AI & Society 40(5): 3891–3907.

75.

Ryan

Isakhanyan

Tekinerdogan

(2023) An interdisciplinary approach to artificial intelligence in agriculture. NJAS: Impact in Agricultural and Life Sciences 95(1): 2168568.

76.

Schmidt

Schutte

Buttigieg

, et al. (2024) Mapping the regulatory landscape for artificial intelligence in health within the European Union. npj Digital Medicine 7(1): 229.

77.

Shahriar

Corradini

Sharif

, et al. (2026) The role of generative artificial intelligence in digital agri-food. Journal of Agriculture and Food Research 20: 101787.

78.

Sigfrids

Leikas

Salo-Pöntinen

, et al. (2023) Human-centricity in AI governance: A systemic approach. Frontiers in Artificial Intelligence 6: 976887.

79.

Smuha

(2019) The EU approach to ethics guidelines for trustworthy artificial intelligence. Computer Law Review International 20(4): 97–106.

80.

Sood

Sharma

Bhardwaj

(2022) Artificial intelligence research in agriculture: A review. Online Information Review 46(6): 1054–1075.

81.

Sparrow

Howard

Degeling

(2021) Managing the risks of artificial intelligence in agriculture. NJAS: Impact in Agricultural and Life Sciences 93(1): 172–196.

82.

Stogiannos

Malik

Kumar

, et al. (2023) Black box no more: A scoping review of AI governance frameworks to guide procurement and adoption of AI in medical imaging and radiotherapy in the UK. The British Journal of Radiology 96(1152): 20221157.

83.

Sullivan

Gemtou

Anastasiou

, et al. (2024) Building trust: A systematic review of the drivers and barriers of agricultural data sharing. Smart Agricultural Technology 8: 100477.

84.

Tagkopoulos

Brown

Liu

, et al. (2022) Special report: AI institute for next generation food systems (AIFS). Computers and Electronics in Agriculture 196: 106819.

85.

Tagkopoulos

Earles

Lemay

, et al. (2024) The AIFS institute: Building a better food system through AI. AI Magazine 45(1): 89–93.

86.

Taher

(2023) Stakeholders’ opinions support the people-process-technology framework for implementing digital transformation in higher education. Technology, Pedagogy and Education 32(5): 555–567.

87.

Taneja

Nair

Joshi

, et al. (2023) Artificial intelligence: Implications for the agri-food sector. Agronomy 13(5): 1397.

88.

Theofanos

Choong

Jensen

(2024) AI use taxonomy: A human-centered approach. Available at: https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=956852.

89.

Tzachor

Devare

King

, et al. (2022) Responsible artificial intelligence in agriculture requires systemic understanding of risks and externalities. Nature Machine Intelligence 4(2): 104–109.

90.

Uddin

Chowdhury

Kabir

(2024) Legal and ethical aspects of deploying artificial intelligence in climate-smart agriculture. AI & Society 39(1): 221–234.

91.

United Nations Educational, Scientific and Cultural Organization (n.d.) Ethics of artificial intelligence. Available at: https://www.unesco.org/en/artificial-intelligence/recommendation-ethics.

92.

Uren

Edwards

(2023) Technology readiness and the organizational journey towards AI adoption: An empirical study. International Journal of Information Management 68: 102588.

93.

van der Burg

Giesbers

Bogaardt

, et al. (2024) Ethical aspects of AI robots for agri-food; a relational approach based on four case studies. AI & Society 39(2): 541–555.

94.

van Hilten

Ryan

Blok

, et al. (2025) Ethical, legal and social aspects (ELSA) for AI: An assessment tool for agri-food. Smart Agricultural Technology 10: 100710.

95.

Wheeler

Cohen

Xiong

, et al. (2020) Sample size for latent Dirichlet allocation of constructed-response items. In: The Annual Meeting of the Psychometric Society, Virtual Conference. Cham: Springer, 263–273.

96.

Wójcik-Gront

Zieniuk

Pawełkowicz

(2024) Harnessing AI-powered genomic research for sustainable crop improvement. Agriculture 14(12): 2299.

97.

Wolfert

Verdouw

, et al. (2017) Big data in smart farming—A review. Agricultural Systems 153: 69–80.

98.

World Bank (2023) Agriculture, forestry, and fishing, value added (% of GDP). Available at: https://data.worldbank.org/indicator/NV.AGR.TOTL.ZS.

99.

World Bank (2025) Employment in agriculture (% of total employment) (modeled ILO estimate). Available at: https://data.worldbank.org/indicator/SL.AGR.EMPL.ZS.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.17 MB

0.00 MB