Exploring the readiness of publicly funded researchers to practice responsible research and innovation in digital agriculture

ABSTRACT

Responsible Research and Innovation (RRI) has been highlighted as necessary for the governance of digital agriculture. The current study explores the readiness of publicly funded researchers in Ireland to engage in RRI activities in digital agriculture. In-depth one-to-one interviews were carried out with 15 scientists and research funders to understand how perceived emergent risks in digital agriculture are being addressed and to reflect on the extent to which responses align with RRI principles. Thematic Analysis identified three themes: (1) Cultural impacts: Addressing unintended consequences of technology; (2) Farm-level impacts: Ensuring user-friendly technologies; and (3) Perceived roles and responsibilities of scientists. There was evident alignment to the values underpinning RRI; however, an integrated RRI approach in digital agriculture may encounter challenges. There is a need to consider how RRI exercises and dimensions are framed, introduced and subsequently supported in science and academia.

KEYWORDS:

• Digital agriculture
• responsible research and innovation
• risk perception
• scientists
• smart farming

Introduction

While technological innovation and digitalisation have been a permanent feature in agriculture for many decades now (Shepherd et al. 2018), the idea of a radical transformation of the agricultural sector is gaining attention with the increased application of ‘smart’ technologies in farming and food production (Rose and Chilvers 2018). The introduction of robotics, Internet of Things, sensors, Artificial Intelligence, drones, and satellites into farming, has led some to argue that the fourth agricultural revolution (‘Agriculture 4.0’) is on the horizon, if not already here (Rose and Chilvers 2018; Shepherd et al. 2018). These technological innovations are designed to solve particular problems; overcome labour shortages, feed a growing population, mitigate environmental impact, improve on-farm efficiency and quality of life and meet societal demands for healthier, safer, more ethical and more transparent food production (Shepherd et al. 2018). However, in providing solutions to these issues, they also shape certain future social realities at the expense of others and may even create new risks and dilemmas (Bronson 2018; Carolan 2020). Different actors will hold different opinions about what the future of agriculture should look like and what role technologies should have in shaping this future. Introducing ‘values-based questioning’ into the decision-making process encourages us to question the type of future society we want and to critically reflect on how our innovations are going to deliver that society (Bronson 2018; Fleming et al. 2018; Glerup, Davies, and Horst 2017). Such discourse-based approaches aim to facilitate diverse actors to debate and discuss desired and undesired outcomes, foster mutual understanding, identify common values and ultimately develop collectively acceptable, or tolerable, decisions (Renn 2008). These principles are promoted under the Responsible Research and Innovation (RRI) approach.

Responsible research and innovation

As a values-based governance framework, RRI aims to support techno-scientific progress in a socially and morally responsible manner (Stilgoe, Owen, and Macnaghten 2013; Von Schomberg 2013). The framework encourages researchers, innovators, policy-makers, and all individuals in key governance positions to be conscious of and responsive to the needs and values of diverse societal actors. The need for responsible innovation in digital agriculture has been highlighted by previous researchers (Rose and Chilvers 2018; Eastwood et al. 2019; Bronson 2018). Whilst acknowledging different approaches to defining RRI, the current paper agrees with previous authors in viewing a process-oriented RRI model (Owen, Macnaghten, and Stilgoe 2012; Stilgoe, Owen, and Macnaghten 2013) as being of particular value for digital agriculture (Rose and Chilvers 2018; Eastwood et al. 2019; Bronson 2018). This model suggests a procedure for RRI centred around four dimensions that guide technology research and innovation.

• Anticipation supports exploration of the possible impacts of research and innovation (including social and ethical impacts), reflection on how research and innovation may shape the future, and engagement in upstream risk assessment.

• Inclusion encourages engagement of diverse voices at an early stage, and on a continuous basis, during the research process, ultimately resulting in the collection of diverse types of knowledge.

• Reflexivity fosters actors to critically reflect upon their own assumptions, values and interests, and actively consider the views and values of others and how they may correspond or conflict with one’s own views.

• Responsivity ensures that actors, as well as institutional structures and policies, take meaningful action in response to insights that emerge during the RRI process so that the research and innovation process is adapted to align with the needs expressed by other actors.

Recent empirical studies have revealed areas of concern with increased digitalisation in agriculture such as growing digital divides, digital exclusion, distribution of power, data ownership, knock-on effects of technological reliance, economic and cultural impacts, and consumer backlash to emergent technologies (Jakku et al. 2019; Regan 2019; Fleming et al. 2018; van der Burg, Bogaardt, and Wolfert 2019; Bronson 2019; Carolan 2020; Small 2017; Fielke, Taylor, and Jakku 2020; Klerkx, Jakku, and Labarthe 2019). RRI is particularly valuable for digital agriculture as there is a need to consider the consequences of technological innovation at different scales including at farm-, community- and societal-level (Rose and Chilvers 2018) and different stakes are evident from the publicly funded research sector, the private AgTech sector, governmental and policy sector, the farming community and broader civil society (Wolfert et al. 2017).

Willingness and readiness of scientists to practice RRI

Embedding RRI necessitates changes in how research and innovation are carried out (Ludwig, Macnaghten, and Pols 2019). Eastwood et al. (2019) refers to ‘RRI readiness’ as the suitability of existing conditions and the capacity of key actors to embed RRI dimensions in research and development projects and initiatives. Successfully embedding RRI also depends on key actors being willing and able to actively practice RRI-aligned activities in their day-to-day roles (Carrier and Gartzlaff 2020). There is an individual attitudinal component to this readiness: Do these actors believe in the value of RRI? Are they motivated to practice RRI activities? There is also an external component to this readiness: Are key actors empowered by their social and cultural environment to practice RRI? Do existing infrastructure, relationships and mechanisms support RRI activities? While RRI is a framework for all actors in positions of power and authority, its impact and relevance for the publicly funded scientific community is particularly apparent as they face increased pressure from funding stipulations, national research policies, and institutional research impact and innovation strategies to meet social, moral and ethical demands, and embrace upstream engagement with diverse audiences (Glerup, Davies, and Horst 2017; Owen et al. 2021). RRI offers a practical means for meeting these demands. In practicing such RRI activities, however, scientists must assume new roles and responsibilities, and embrace new ways ‘to do research’ (Stilgoe, Owen, and Macnaghten 2013; Ludwig, Macnaghten, and Pols 2019). There has been limited research exploring the actions scientists take to materialise RRI in their day-to-day roles (Glerup, Davies, and Horst 2017). Studies suggest that whilst scientists show commitment to the ideals promoted by frameworks such as RRI, there is uncertainty, and sometimes ambivalence, around how to implement practices that actually promote these ideals (Glerup, Davies, and Horst 2017; Frankel 2015; Felt, Fochler, and Sigl 2018). Specific to exploring RRI in digital agriculture, previous authors (Rose and Chilvers 2018; Bronson 2018; Eastwood et al. 2019) have made suggestions about indicators which are reflective of each of the four RRI dimensions, which often tend to be participatory tools for inclusion.

It has been argued that inclusion acts as the central lever for practicing RRI, as anticipation, reflexivity and responsiveness are all dependent on first embedding a process to engage and capture diverse voices (Rose and Chilvers 2018). Acts of engagement are indeed a central mechanism for RRI as they support the required collective deliberation amongst all relevant actors to ensure research and innovation is responsive to societal needs (Lorono-Leturiondo and Davies 2018). However, we know from earlier and related science governance movements – (e.g. Public Understanding of Science; ELSI: Ethical, Legal and Social Issues of science) that engagement is not always a role fully embraced by publicly funded scientists (Li et al. 2015). Some studies show scientists as unwilling to engage with non-academic audiences because they doubt the technical capacity of ‘lay audiences’, while other studies highlight cultural and organisational constraints (Li et al. 2015; Horst 2013). Where RRI practices are concerned, it is likely that tensions may exist in the types of roles and responsibilities which scientists are willing and able to assume (Henkel 2005). Accordingly, the current paper takes the perspective that it is important to not just explore how social and ethical issues are currently being addressed within digital agriculture research, but also to explore the readiness – including the willingness, inclinations and underlying attitudes – of the scientists charged with addressing them. The motivations and beliefs of these scientists will significantly influence the success of any future exercises to embed formalised RRI activities in digital agriculture. Indeed, in considering the dimension of ‘reflexivity’, one can argue that it is this, and not inclusion, which is the core lever powering the RRI framework (Felt, Fochler, and Sigl 2018). Scientists require this skill and mind-set to ensure that all other dimensions are undertaken with a genuine willingness to anticipate, engage, reflect upon and respond to the concerns, priorities and needs of other actors.

Context: publicly funded digital agriculture research in Ireland

In Ireland, agri-food represents one of the most important indigenous sectors, primarily export-led and accounting for 7.6% of all employment in the country. Ireland’s farming tradition is facing macro-level challenges including climate change, economic and environmental sustainability, rural decline, and Brexit uncertainty. Debates are intensifying regarding the future shape the sector will, and should, assume. Farming is intrinsically linked to the social and cultural rural fabric of Ireland, with predominantly small, family-owned farms comprising the sector. With this family farming tradition, comes decision-making which is influenced not just by economic considerations, but also social and cultural considerations. This is the value-laden backdrop to the technology-driven revolution forecast for the Irish agri-food sector in the coming decades (Teagasc 2016). Digitalisation in Irish farming brings unique challenges, primary of which are concerns of amplified inequities and divides between smaller and bigger farms, sectors and individual farmers (e.g. age and digital literacy disparities) and changing power dynamics brought about by the entry of new actors (e.g. tech companies and investment actors) into the agricultural landscape (Regan 2019).

Digital agriculture is a rapidly emerging and developing sector in Ireland with increased public funding devoted to research and innovation in this sector in recent years (Regan 2019). In the last number of years, there has been increased attention on RRI as a governance framework in Ireland, primarily within public institutions funding research and innovation in Ireland. The European Union, the Department of Agriculture, Food and the Marine (DAFM), and Science Foundation Ireland (SFI) are the primary public funders of digital agriculture research in Ireland and each of these funders have incorporated RRI into their programme structures. The position of RRI within the European Union’s Framework Programmes is well-discussed (Von Schomberg 2013). SFI and DAFM require all funded research centres, programmes and proposals to report on RRI elements with specific indicators such as public engagement and science education along with open science, gender and ethics. A specific focus on outreach aligns to broader increased policy interest in Ireland towards citizen participation movements in response to local and regional technological conflicts (Hennen and Nierlin 2015). The main formalisation of RRI in Irish digital agriculture research has been through activity at the funding level as well as an endorsement by universities and research institutes of the National Policy Statement on Ensuring Research Integrity in Ireland. However, it is unknown to what extent, and how, this macro-level emphasis on RRI has seeped into the mind-sets and behaviours of Irish agricultural researchers. While other countries have explored the practical application of RRI in digital agriculture research including in New Zealand (Eastwood et al. 2019) and North America (Bronson 2019), no similar exploration has taken place in an Irish context. The current study explores current practices and attitudes in the Irish scientific community towards dealing with the social, ethical and moral risks that could emerge with the rise of digital technologies in agriculture. It reflects on how those practices and attitudes align to RRI principles. In doing this, it aims to provide insights on (1) the extent to which researchers are likely to be motivated and positively inclined towards RRI activities; and (2) identify what existing practices could be leveraged, and what changes might be needed, to better align to an RRI framework in publicly funded digital agriculture research in Ireland.

Materials and methods

Sample

One-to-one, in-depth semi-structured interviews were carried out with individuals in the Irish research community. A purposeful sampling strategy was used to recruit participants who had (a) a strong strategic overview of, and experience in, publicly funded research and innovation in Smart Farming and (b) came from different disciplinary backgrounds. Potential candidates were identified by reviewing agenda and participant lists of conferences/events and research/policy working groups, and searching the websites of relevant research groups in universities and research institutes in Ireland. Purposive selection of participants was based on examining prospective participants’ scientific activity (research projects, papers, conferences) and impact/dissemination activity (media & online output, policy/commercial engagement). An invitation to participate in the study was issued via e-mail and followed up with a phone-call. The sample was continuously monitored during data collection and targeted, purposeful sampling was employed to ensure sufficient spread of participants across disciplinary backgrounds.

Interviews were carried out with 15 individuals¹ from the scientific community. The sample comprised 13 scientists from diverse academic backgrounds including natural sciences (6 participants from horticulture, crops, animal science and agricultural engineering), information and communication technologies (3 participants from computer sciences and spatial analysis) and social and applied sciences (4 participants from agri-business, geography and sociology). The scientists came from 8 different research institutes geographically distributed across the east, south and west of Ireland. An additional 2 interviews were carried out with representatives from the two main research funding organisations funding digital agriculture research in Ireland. All interviewees were at relatively advanced / senior stages of their career, with all of the 13 scientists acting currently or previously as Principal Investigators on national or international research projects. The sample included 4 females and 11 males. The decision to close the data collection after the 15th interview was guided by a number of good practice recommendations relating to sample size determination in qualitative research (Vasileiou et al. 2018; Morse 2000). Theoretical saturation was determined by analysing interview data concurrent to data collection; it was judged that by the 15th interview, few new theoretical insights were emerging pertaining to the core research question. The focused nature of the research aims, the depth of the data collected in each interview and the diversity in participants’ disciplinary research backgrounds facilitated sufficiently rich and meaningful data for analysis.

Semi-structured interviews

The interview schedule was first piloted with one social scientist and one ICT scientist. The findings from the pilot study did not significantly alter the interview schedule content, although question order was restructured to enhance interview flow. An internal organisational review of the study took place and ethical guidelines were followed. Participants were provided with an information sheet and all aspects of confidentiality, anonymity and right to withdraw from the study were explained fully to participants. Following the opportunity to read the information sheet and ask questions, all participants were asked to sign an informed consent form. Following the interview, participants were debriefed and provided with more information on the purpose of the study. They were also offered the opportunity to receive a findings summary when it became available. All data (audio recordings and transcripts) were securely stored and only accessible by the research team.

Participants were invited to choose a location for the interview; for all participants, this was their place of work. Interviews lasted between 50 and 70 min and were conducted face-to-face during the period of April-June 2018. An interview schedule was used to guide the interview (See the Appendix). The research interview began with participants being asked to describe their interpretation of the term ‘smart farming’. They were then presented with a written definition and an image depicting a ‘smart farm of the future’ (depicting survey drones, a fleet of agri-bots, smart tractors, texting cows, and farming data). Participants were asked to ‘think aloud’ and discuss whether they agreed or disagreed with these materials. A semi-structured interview guide was then used to facilitate participants’ discussions on the risks and benefits for smart farming research and innovation in Ireland and their thoughts on future developments. For any risks identified by participants, they were further probed as to how they currently, or would like to, respond to those risks through their own work. Finally, participants were asked about the manner in which they currently carry out research (the funding process, the types of projects they work on, who is involved in the projects; and how they translate their research findings). It was decided not to introduce the concept of RRI to the participants during the interviews. The current study was interested in the current practices and attitudes of scientists with a view to exploring their ‘readiness’ and willingness to embrace RRI. There was a concern that introducing the RRI concept and terminology could detract from this aim. Previous research has found that scientists have responded to RRI terminology with uncertainty and perceived irrelevance to their work, viewing the terminology to reflect a policy-level concept far removed from the work in their offices and laboratories (Felt, Fochler, and Sigl 2018; Glerup, Davies, and Horst 2017). Although it is an interesting and valuable line of inquiry to explore scientists’ perceptions of the RRI framework itself, it was not the primary aim of the current research.

Qualitative analysis

Interviews were audio-recorded and transcribed verbatim. At the request of one participant, written notes were taken in lieu of an audio recording. QSR NVivo 10 assisted in the organisation and management of the analysis. All data underwent an inductive thematic analysis (Braun and Clarke 2006). Familiarisation with the data took place by first actively reading each transcript before proceeding to data analysis. Inductive coding took place to examine in an in-depth manner, the different perspectives and practices discussed by the participants; the coding was grounded in the interview data. Similar data-driven codes were merged to develop coherent themes. Verbatim quotes illustrate themes and sub-themes: quotes were selected for their illustrative ability to reflect the essence of each theme (Braun and Clarke 2006).

Findings

By exploring the practices and attitudes of the scientific community, the data analysis revealed how perceived emergent risks² are being addressed in digital agriculture research and innovation in Ireland. Three central themes were developed. The first theme reflects participants’ concerns about the longer-term, unanticipated social, ethical and moral consequences which could emerge; participants discussed actions to be taken including the development of policy-led strategic goals, multi-actor collaboration, and formalised anticipatory exercises within research and innovation. The second theme reflects participants’ concerns about the more immediate, negative socio-economic impacts which could be felt at farm level with the development of irrelevant or non-valuable smart technologies; participants discussed responses in the form of farmer engagement in research and innovation, and public-private collaborations. The third theme discusses perceived roles and responsibilities of the scientist as a motivating factor which strongly shaped participants’ views and practices towards methods for addressing the aforementioned issues. Given that the current study was interested in understanding the readiness of the Irish scientific community to practice RRI in digital agriculture, this section reflects on the alignment of the observed practices and views with the RRI framework and the dimensions of anticipation, inclusivity, reflexivity and responsiveness.

Cultural impacts: addressing unintended consequences of technology

As a major theme, participants engaged in a process of reflection and visioning of the potential risks arising from future smart farming scenarios in the form of ‘unintended consequences’, a concept which resonated strongly with participants. Negative unintended consequences were viewed as arising from widespread and cumulative adoption of technologies; were social, moral and ethical in nature; and their impacts were felt at a broad societal and cultural level (as opposed to the more immediate and direct effects which adoption of single technologies can have at, for example, individual farm level). Echoing findings of other studies (Jakku et al. 2019; Fleming et al. 2018; van der Burg, Bogaardt, and Wolfert 2019; Bronson 2019; Carolan 2020; Small 2017), potential unintended consequences included: increased digital divides, technological unemployment, a loss of traditional farming culture and tacit knowledge/skills, changes to farmer identity, the demise of certain types of farms, increased farmer isolation, reduced human-animal interactions (with hypothesised impacts on animal welfare), consumer or societal backlash, data-driven power imbalances, and data-driven threats to individual privacy and rights. Discussion of these negative unintended consequences prompted participants to highlight the need to actively reflect on how digital technologies may (negatively) shape the future.

I think it’s hugely important to look at this from a much more holistic process. I think the positives and negatives of the financial, social, cultural implications need to be seen. (Scientist 11, social and applied sciences)

This quote shows a willingness to engage in the type of anticipatory thinking encouraged under RRI. While all participants agreed that unintended consequences were a reality of new technologies, participants held different views around how to address them. Some reflected on unintended consequences that could arise from the recent and rapid growth of the private AgTech industry both in Ireland, and globally. Participants referenced a strong presence of ICT industry in Ireland which was viewed favourably for further developing digital agriculture research and innovation, but also cautioned the extent to which this sector would understand and be responsive to the requirements of Irish agriculture. ‘Joined-up thinking’ and collaborative efforts were viewed as particularly important to ensure that the trajectory of digital agriculture would not become detached from the priorities and goals set out for the national development of Irish agriculture.

You know there’s all sorts of bits and pieces of technology being sold. Farmers have been assailed from all different directions, with different products … So what we have here in Ireland are those small Irish technology companies as well as the multinationals. But they have … to come together. And we as publicly-funded researchers can take a role in bringing them together … we have an important role to play in terms of that public good role … in supporting and helping guide that development towards, what are the real needs of farmers here. (Scientist 9, natural sciences)

The quote from this participant illustrates how actors in the research community are thinking about where their responsibilities as publicly funded researchers lie in ensuring a strategic approach. Participants from the research funding agencies were particularly vocal on highlighting the value of a foresight or prioritisation exercise which could feed into the development of a national policy on digital agriculture to guide research and innovation. A strategic overview of the priorities and needs of Irish agriculture would help develop clear policy objectives, a common vision and joint goals to work towards and this would help support multi-actor collaborations. The need for a policy-driven strategic overview has previously been identified as a good governance practice for guiding research and innovation associated with other major transformations within the agricultural sector, such as the bioeconomy (Asveld, Ganzevles, and Osseweijer 2015). Setting out a national strategic overview could help provide the centralised coherence needed to responsibly and effectively develop digital agriculture and to identify and avoid unanticipated impacts.

A number of participants talked about the practices which could be employed to formalise reflection of unanticipated consequences within research projects. Suggestions included for example, future-casting processes (foresight exercises, road mapping, scenario testing); multi-actor projects and inter-disciplinary projects (with social scientists embedded); and public engagement (‘talking to communities’). Engagement was viewed as a way to ensure relevant actors have a say in decision-making related to the governance and direction of technology development in agriculture.

Take research prioritisation exercises, it definitely makes sense for the public to have an opportunity to influence that … in farming, there are going to be controversial things. Things like cloning research, stem cell research, it makes sense that the public would get to have some input on – are Ireland going to do those things? Do we think that’s okay? (Scientist 3, information & communication technologies)

Participants such as this scientist spoke of the value of formalised engagement exercises that make the research and innovation process more transparent and more democratic. Such thinking reflects the principles of the RRI framework. Some participants also made specific reference to the value that could be achieved by thinking through different scenarios from different perspectives. Such formalised structures serve to question current assumptions and ideologies, as illustrated in the following quote of one of the social scientists:

I think the model where science and extension are in the same place is quite good because you have more interaction which allows you to be a bit more reflective about the role of technologies. There’s often an idea that ‘technology equals progress equals profit’ so technology is just seen as good. But what does that actually mean, and is it actually good for farmers? So just being a bit more reflective – is technology always good? (Scientist 1, social & applied sciences)

Such sentiments evidenced participants’ willingness to engage in deliberative decision-making, in line with the RRI dimension of reflexivity (Felt, Fochler, and Sigl 2018). And the aforementioned mechanisms suggested by participants also support an alignment with RRI as they correspond to previously suggested indicators reflective of the RRI dimensions of anticipation and inclusion (Rose and Chilvers 2018; Eastwood et al. 2019; Bronson 2018). Thus, when it comes to reflecting on the longer-term social, moral and ethical impacts of digital agriculture research and innovation, it was clear that the attitudes and suggested responses expressed by at least some in the sample corresponded well to the RRI framework. However, some contradictions to this pattern were also noted. A minority of scientists in the study held some cautionary views about the act of formalising anticipatory thinking within research and innovation projects. Views were expressed that unless they were carefully managed, such exercises could portray negative connotations, disproportionately amplify risks and hamper innovation. Some also questioned the futility of ‘box-ticking exercises’ trying to ‘predict the future’, as characterised by one natural scientist (Scientist 6), or ‘crystal ball gazing’ as characterised by another (Scientist 13). These quotes demonstrate how some scientists view the aim of these exercises to be the development of accurate future scenarios, and perhaps did not see the broader value in the deliberative mechanisms facilitated by such processes. From an RRI perspective, a primary marker of the success of such exercises is the fostering of more reflexive thinking of those involved (Felt, Fochler, and Sigl 2018). The concerns hint at a possible perceptual challenge in gaining trust and buy-in amongst all in the scientific community for RRI exercises specifically aimed at reflecting upon longer-term, often unknown, consequences of technologies. Previous research has found that scientists are unfamiliar with the concept of RRI, or feel it is a far-removed policy concept with little application or relevance to their day-to-day roles (Felt, Fochler, and Sigl 2018; Glerup, Davies, and Horst 2017). The current findings reinforce the need to carefully reflect on how RRI exercises are framed and presented to the scientific community.

Farm-level impacts: ensuring user-friendly technologies

A second major theme discussed by participants reflected the perceived risks of technologies being developed for farmers, and being invested in by farmers, which do not account for farmers’ needs. Digital technologies were viewed in principle to bring significant economic and social benefit to farmers. However, concern was raised that the manner in which some technologies are developed may mean that farmers fail to see a return for their investments. A small number of participants also cautioned that technologies should not be developed with only specific types of farmers in mind, leading to the exclusion of segments of the Irish farming community, particularly older farmers and smaller farms. This has previously been noted in the literature with respect to the values driving smart farming in Canada and the extent to which smaller farms are being reflected in new technological trajectories (Bronson 2018). Participants spoke of how the smart farming innovation trajectory had taken off at speed within Ireland. Some participants noted a saturation of technologies on the market, and raised concerns about a lack of proper validation.

There's a guy up the road put in a really top of the range parlour recently and he spent so much money on it and he was telling me one day about all the technology and all the stuff he gets on the phone. And I asked him how he's using it and he said he actually is not, he switched it off because it was just hitting him with too much information. So how do we change that? We change it by providing him with information that’s tailored to his needs. Just provide him with information that’s timely and through the appropriate medium and that’s part of our job in our new research project is to get a better understanding of that. (Scientist 6, natural sciences)

Aligned with the principles of RRI, participants such as this scientist talked about how they were responding to these challenges to try and change the direction of this innovation trajectory – to ensure that technologies being brought to market were validated and proven to be of benefit for Irish farmers, and for Irish agriculture.

There was strong support amongst the majority of participants for a partnership approach between publicly funded research and the private AgTech industry. With a strong agricultural base and a strong ICT community, participants perceived good opportunities in Ireland for collaboration. A number of the scientists discussed their involvement in jointly funded demonstration farms or large public-private research projects and the role that they played in these projects, in verifying and validating the value of technologies for Irish farmers.

In their responses to the risk of irrelevant or even damaging technologies being developed for Irish farmers, all participants strongly believed that farmers – viewed as the primary end-users of digital technologies – should be involved in the research and innovation process. Common across participants was a desire to ‘listen to the farmer’ and to understand their needs, as a user of technology.

I think when developing technology we have to keep in mind who’s going to use this technology and for what purpose. Part of the process of development has to include this loop that says well who’s going to use this? How are they going to use it? What are going to be their opportunities? What are going to be any difficulties that they are going to experience? What are the implications for them? That process is not complete until the end user is incorporated. (Scientist 8, social & applied sciences)

There was significant evidence of current practices where farmers were already being included in research projects via formalised ‘engagement mechanisms’. Participants identified a range of mechanisms and practices: stakeholder groups; farmer technology groups; operational groups; knowledge transfer groups; design thinking; co-creation; on-farm pilot studies; demonstration farms; farm open days; farmer conferences; multi-actor projects; interdisciplinary research and the role of social science. These practices strongly mirror the indicators proposed for the ‘inclusion’ dimension of the RRI framework (Rose and Chilvers 2018; Eastwood et al. 2019; Bronson 2018), although of note, within this theme – inclusion was very much centred on engaging one particular type of stakeholder – the end-user, the farmer. This highlights the need to reflect on how ‘rights holders’ are identified within digital agriculture innovation (Bronson 2018), and in particular the extent to which actors traditionally outside of immediate agricultural networks are engaged (including for example, civil society).

With respect to farmer engagement, the primary role imagined for the farmer was to become involved in the design process so as to test ideas and technologies, and to provide feedback on the usability, applicability and relevance of the technology under development. Participants strongly advocated for an agile, iterative approach to technology development with incremental changes to design features and deployment based on continuous feedback from the farmer until the technology is developed to best suit the needs of the farmer. Early stage validation was viewed as important for the ultimate success of the technology. This approach was viewed to allow the required space for trial and error, and to see how farmers work and interact with the technology. The ultimate aim of this farmer engagement was to achieve consensus on how best to design and deploy technology; to involve the farmer earlier in the process to build trust in the technology and secure buy in; to develop a more valuable technology for the end-user, one compatible with their needs; and ultimately, to ensure successful technology adoption and engagement.

You have to include farmers as end users because they will come up with new ideas or uses but secondly you're starting to get buy-in from the beginning, which is very important if there's something new coming. Our programme is focusing on precision agriculture and we have been trying to engage farmers in that because I think more than most other technologies you have a huge divergence in opinions and use. (Scientist 5, natural sciences)

Participants such as this scientist discussed how interactions and engagements with the end-user, the farmer, helped them to trigger different ways of thinking and reduce the risk of making incorrect assumptions about their needs; sentiments which are reflective of RRI. Importance was placed on a two-way process of learning with participants indicating that they were able to learn from the farmer about their specific needs and desires with respect to the technology under design. A few participants also referenced the value in taking the engagement process a step further and having the farmer involved as a co-creator of a new technology.

Farmers see new possibilities on how to do things, it’s easy for us as academics to sit at desks and think of ideas, it’s only when you’re out there that you start to see the kind of innovation that comes about from having the farmer as part of the conversation … We have to empower the farmers … get that kind of space to somehow swap ideas and, and to learn from each other. (Scientist 4, information & communication technologies)

This sentiment is perhaps most in fitting with the principles of RRI. Inclusion of the farmer in the process is viewed to involve a deeper two-way learning process; no longer just learning about the needs of the farmer in relation to a specific technology, but to have the farmer contribute their tacit knowledge so to identify the problems that need to be tackled and co-create new ideas or approaches.

Perceived roles and responsibilities of scientists

A strong cross-cutting and major theme, participants in the current study expressed particular beliefs about the roles and responsibilities a scientist should fulfil, which appeared to shape their attitudes towards activities aimed at anticipating long-term impacts or involving different actors in research and innovation. From the perspective of the research funders interviewed in this study, carrying out research in an inclusive and anticipatory manner was viewed as an increasingly required component of the researcher’s portfolio. They made specific reference, for example, to the need for multi-disciplinary, multi-actor projects and stakeholder/public engagement.

We fund research projects now and we give them marks for their public engagement approach. And you can lose funding because you don’t engage. We fund in some research areas where they are out talking to local communities, decades before something will be deployed on a commercial level. The tech development is nowhere near commercialisation yet but they’re starting to talk to communities already … and opening up that discussion. (Research funder 2)

Views such as this reflect how macro-level changes in science governance have become embedded at the funding level, and in the mind-sets of the researcher funders, which are subsequently having an impact at the micro-level, bringing new expectations on the roles and responsibilities of the researcher (Henkel 2005).

For the scientists however, tensions existed between what they perceived their role as a scientist to be, and where, and how, these formalised exercises aligned with that role. Individual beliefs about one’s professional identity – termed ‘academic identity’ with respect to scientists and academics – has long been shown to influence willingness to assume new roles and responsibilities (Henkel 2005). On the one hand, participants felt it was their duty to ensure that the development of digital agriculture was underpinned by scientific evidence and an unbiased weighing up of the risks and benefits. As illustrated by Scientist 8, social & applied sciences: ‘I think we have a duty and a responsibility to anticipate. If the scientists don’t identify the potential unintended consequences, well who will?’ Conceptually, scientists value the idea of carrying out ‘responsible’ research, as previous research has also found (Felt, Fochler, and Sigl 2018; Glerup, Davies, and Horst 2017). Participants discussed their role in providing unbiased and independent assessments of technologies being brought to market by the AgTech industry, and working with private industry so as to ensure that the needs of farmers, and of the Irish agricultural sector, were being appropriately met.

The private sector is racing ahead. There’s many, many companies … who are developing all sorts of machines and bits of equipment … which they’re selling to farmers. So our role can be to evaluate these things, see what works well, what’s of good value for farmers. (Scientist 9, natural sciences)

For scientists such as this, they see it as part of their role to ensure that technologies are not brought to market by private industry in the absence of any validation studies, impact assessments or engagement with the farming and agricultural community. It was important for these scientists that values such as scientific ‘truth’ and rigour were embedded in digital agriculture research and innovation. This aligns with previous research which has found academics and scientists tend to use values such as a commitment to quality, rigour, autonomy and freedom to inform their views on the expected roles and responsibilities of their profession (Degn 2018; Henkel 2005). However, research has highlighted that scientists’ also believe that an important part of their role is to demonstrate expertise, and that this is best exemplified by demonstrating excellence in one’s specific discipline (Schikowitz 2020), and in the current study, this is where conflicts appeared to arise. Participants discussed how the perceived roles and responsibilities of the scientist are largely shaped by their beliefs around what types of knowledge should drive the research and innovation process, with scientists often favouring scientific knowledge over lay knowledge.

It’s much easier to have a team in a lab, who are, well ‘they know best’. By actually taking it out to these so called ‘lay-people’ … you know it’s not always the case that these professionals want to work with them and involve them, in their pipedream. There’s a cost and there’s a time delay. And they may not always hear what they want to hear. (Scientist 8, social & applied sciences)

Quotes such as this are reflective of the commitment that scientists often feel to ensuring ‘scientific excellence’ above all else in their jobs (Degn 2018; Schikowitz 2020). Reflexive thinking is not viewed as easily undertaken within research settings which still maintain academic disciplinary silos, and where scientific expertise still ultimately defines the identity of a scientist:

As individual scientists we have a very strong ethical driver to think about the potential consequences. I guess we being computer scientists lean towards the engineering – our first question is can we do it? Rather than, should you do it? That’s inevitable … . (Scientist 2, information & communication technologies)

A small number of participants talked about the role of inter-disciplinary research as a solution to over-come the individual disciplinary biases which may prevent scientists from reflexively thinking about technological development. A particular topic of discussion was the role of social scientists working alongside other scientists and technology developers, however, it was noted that this approach was not without its challenges. Participants cautioned how social science strands are positioned within research projects and the extent to which they have the power to shape the direction of technological and scientific innovation. Speaking from experience, one social scientist indicated the challenges they had encountered in adopting an anticipatory role in inter-disciplinary research projects:

It can be a difficult position for social scientists to be in sometimes because they are the ones expected to solve the ethical issues or else they are the troublemakers bringing up these ethical issues. (Scientist 1, social & applied sciences)

Quotes such as this show the challenges which can come with inter-disciplinary research, often driven from differing values and expectations. It is worth noting as a final sentiment raised by participants, and which has previously been highlighted as a barrier for researchers to adopt new roles (Ludwig, Macnaghten, and Pols 2019; Owen et al. 2021): the role of institutional and cultural pressures and supports. While supportive of anticipatory and inclusion practices, the realities of the job often prevent scientists from engaging in such practices, with one IT scientist highlighting the need for greater institutional support for researchers:

It should come down to the researcher but a lot of us would see stakeholders as a burden … We put a lot of attention on the idea of interaction with stakeholders but it’s really at the beginning and end. It’s not throughout the process where it should be. We are all very busy people and interacting with stakeholders is time-consuming. The responsibility lies with the researcher but they need to be fully supported by the funders and the administrators in order to achieve that. (Scientist 2, information & communication technologies)

This sentiment highlights how RRI for individual scientists can be fostered or hampered by the institutional supports and cultures in place.

Discussion

The current study set out to explore how the publicly funded scientific community in Ireland are already thinking about and responding to social, ethical and moral risks which may arise as this area continues to grow. The attitudes and practices reflected by participants in the current study show some alignment with the dimensions of the RRI framework although a number of challenges are also evident. This discussion section considers what this means for if and how RRI activities may be realised in publicly funded research in digital agriculture.

When responding to unanticipated consequences of digital technologies in agriculture, some participants expressed attitudes which aligned to the RRI framework – they talked about the need for formalised mechanisms for anticipation exercises within research projects; the need for inclusion of diverse voices in decision-making; and the importance of reflexive thinking within science. That said, few participants offered practical examples of actions being currently undertaken to address longer-term social and ethical impacts. It was also apparent that some participants held reservations about formalised processes for reflecting on unanticipated consequences. This is not surprising – research previously exploring scientists’ views of RRI unearthed similar worries regarding loss of autonomy, and the futility of anticipating future research outcomes (Carrier and Gartzlaff 2020). Without a clear understanding of the goals of RRI exercises, there is the potential for scientists to view them as excessively cautious, even hinging on alarmist, or disregard them as ‘pointless exercises’ in the belief that it is impossible to predict the future. This perceptual challenge needs to be addressed by clarifying the goals of such exercises for scientists and demonstrating relevancy (Felt, Fochler, and Sigl 2018; Glerup, Davies, and Horst 2017). Furthermore, as suggested by Glerup, Davies, and Horst (2017), there is a need to ensure that the language and framing used around practices of responsibility should be shaped by how scientists themselves discuss these issues, rather than enforcing policy-driven language which may be viewed with irrelevance or suspicion. A valuable line of future research would be to consider the manner in which to best introduce aspects of the RRI concept to the scientific community. The current study is grounded in the 4-dimension RRI framework (Owen, Macnaghten, and Stilgoe 2012; Stilgoe, Owen, and Macnaghten 2013). For RRI scholars, this remains a particularly useful theoretical framework to strategically assess the practices and governance of science and technology development. However, it is less clear whether similar value is to be gained by introducing such frameworks and theoretical concepts into the vocabulary and practices of those scientists working within everyday digital agriculture research. In considering how best to embed RRI principles into everyday practices, it appears that the lever which would most likely be grasped and welcomed by the participants in the current study was that of ‘inclusion’.

The dimension of inclusion has been argued as a key lever which can facilitate the practice of RRI (Rose and Chilvers 2018). The findings revealed an apparently normalised practice of engaging farmers in research and innovation; with participants referencing a range of platforms already being used within research and innovation circles. Furthermore, private sector engagement, which is also viewed as an important aspect of inclusion (Eastwood et al. 2019), was apparent in the current study. This is likely in part as a result of recent trends in Irish research and innovation towards increased citizen participation to avoid technological backlash and also an emphasis within the sector towards ensuring that research is translated into commercial impact (Hennen and Nierlin 2015). Rose and Chilvers (2018) previously suggested using mapping methodologies to support the identification of existing spaces of participation in agricultural research that could be better utilised rather than, or along with, embedding new mechanisms for inclusion within research. The current study indicates that there are quite a number of platforms and channels which could be leveraged to further facilitate inclusion within research and innovation in digital agriculture in Ireland. There is a need, however, to consider the goals currently served by these engagement practices. In related research, it has been questioned to what extent stakeholder inclusion genuinely opens up the innovation process (van Mierlo, Beers, and Hoes 2020). Similarly, in the current study it can be queried whether inclusion in all cases leads to meaningful collaborative action. The majority of the narrative around farmer engagement centred on issues of usability and design and tended to happen at the level of individual technology design. It has previously been noted that research within digital agriculture follows a more top-down approach and inclusion, when it happens, is generally on a project- or technology-specific basis where the aim is to adapt a specific innovation so as to increase stakeholder acceptance (Rose and Chilvers 2018; Chilvers and Kearnes 2016). With RRI, the aim of inclusion is not always to achieve consensus, but to be open to receiving critical feedback (Stilgoe, Owen, and Macnaghten 2013; Von Schomberg 2013). If leveraging existing engagement platforms to support RRI practices, careful consideration once again needs to be given to ensuring clarity over the goals of such exercises.

Previous research highlighted how scientists view ‘stakeholders’ as a more important actor for two-way engagement than the ‘general public’, as stakeholders are perceived to have more realistic views about practical requirements and societal demands (Carrier and Gartzlaff 2020). Thus, it was perhaps not unsurprising to find in the current study also that while there was some support for engaging the general public, the narrative was overwhelmingly centred on the need for engagement with stakeholders such as farmers and agri-tech. Under RRI, inclusion requires the collection of ‘diverse voices’ (Stilgoe, Owen, and Macnaghten 2013; Von Schomberg 2013) while Bronson (2018) speaks of the need for ‘wide involvement of rights holders’. The extent to which civil society and citizens are engaged in digital agriculture, and the extent to which scientists are willing to engage this group of actors, is worth reflecting on. It’s a democratic right for the public to have a say in decisions that affect them, and their contribution ensures a more rounded, diverse collection of knowledge resulting in decisions being made which have a better chance of conforming to local needs and priorities (Li et al. 2015). It is well documented that traditionally scientists have struggled with the idea of engaging the general public in the research process, because of personal beliefs or due to organisational constraints (Li et al. 2015; Horst 2013; Carrier and Gartzlaff 2020). There is a need to consider how this resistance to such engagement may play out specifically in digital agriculture, as there has been little scientific inquiry in this area to date, despite emerging evidence to suggest it will be of increasing importance for the future integration and acceptance of these technologies at market level. A recent survey of the German general population found relatively high public support for funding of research for digital farming livestock technologies, but also identified a growing disconnect between the general public and the agricultural sector which led to emotionally driven concerns about these technologies (Pfeiffer, Gabriel, and Gandorfer 2021). Given that societal reactions to modern livestock practices often generate much debate (Boogaardt et al. 2011), the means and extent to which the general public are engaged in both conversations and innovation processes is an area of inquiry deserving of more reflection.

Perhaps the most important ingredient for embedding RRI in practice is to instil in researchers, a commitment to do research in a way that aligns to RRI values. The primary means by which to ensure an integrated RRI approach is to ensure that there is a commitment to reflexivity by all involved, and, that opportunities for reflexivity are present, for example, during anticipatory and inclusion exercises (Felt, Fochler, and Sigl 2018). Where reflexivity is present, researchers are actively assessing their own motivations and assumptions, reflecting on the motivations and values of others, acknowledging tensions, and taking action to respond to those tensions (Stilgoe, Owen, and Macnaghten 2013; Von Schomberg 2013). This begs the question of how we can best support and cultivate a norm of RRI in science and academia whereby scientists are willing, and able, to commit to these practices. The current study found a discrepancy in terms of the roles and responsibilities expected of scientists under an RRI framework, and the roles and responsibilities which scientists themselves perceive to be most highly valued in their day-to-day work. These tensions have also been identified in related literature exploring scientists’ views of RRI (Schikowitz 2020; Carrier and Gartzlaff 2020). Echoing the findings in the current study, Glerup, Davies, and Horst (2017) previously highlighted how scientists’ reacted to the RRI framework as an imposition placed on scientists from external actors, who were not aware of scientists’ day-to-day realities. Previous research indicates that scientists do not perceive that the practices reflective of RRI (e.g. engagement) are institutionally or culturally rewarded or supported (Ludwig, Macnaghten, and Pols 2019; Owen, Forsberg, and Shelley-Egan 2019). There is a clear need to focus on responses and changes required to support the ‘organisational institutionalisation’ of RRI within digital agriculture research and innovation (Owen et al. 2021). There is value in particularly understanding how organisations may support researchers at all career levels to practice RRI activities. The current study interviewed key informants who were all at relatively advanced levels of their career. At the same time, it is often the case that early career researchers determine the extent to which RRI principles will actually be embedded in day-to-day research activities. Exploring the RRI views and readiness of researchers across different career stages, and the organisational supports they require, would be a valuable area of future research.

Conclusion

Based on the insights gleaned from the current study, and mirroring findings from the wider literature (Carrier and Gartzlaff 2020), there is evidence to tentatively suggest that the attitudes and current practices expressed signify a scientific community who are welcoming of the values underlying the mechanisms and dimensions proposed by the RRI framework. However, as has been noted also in previous research with scientists in many different settings (Carrier and Gartzlaff 2020; Schikowitz 2020) – both individual worries and organisational barriers remain for scientists to meaningfully practice RRI. The current study fully concurs with the concise synopsis from Carrier and Gartzlaff (2020, 166) describing the need to overcome resistance amongst researchers in Europe with respect to RRI: ‘There is good will, and it should not be spoiled.’ In order to ensure innovations in digital agriculture are societal successes – all actors need to be meaningfully engaged in the innovation process. Capacity building for ‘RRI readiness’ (Eastwood et al. 2019) will require a mind-set change for some scientists that will only be brought about by addressing the current tensions that exist between the perceived, valued, and rewarded roles in the science profession. Securing actual buy-in and embedding the RRI dimensions in a genuine and coherent manner will depend on a number of actions to motivate and support scientists to want to commit to new roles and responsibilities. One of the key reflections from this paper is to carefully consider how RRI exercises and dimensions are framed and introduced to the scientific community, as this can influence their willingness and ability to engage in such activities.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Áine Regan

Áine Regan works as a Research Officer (Social Science) with Teagasc, the Agriculture and Food Development Authority of Ireland. Her research areas are behaviour change and risk governance in food and agriculture. She uses behaviour change models to develop evidence-based and societally acceptable strategies for behavioural challenges in food and agriculture. Áine is interested in how risks are perceived, managed and communicated in food and agriculture, particularly where new technologies are at play. Áine has an MSc in Health Psychology (National University of Ireland, Galway) and a PhD specialising in the perception and communication of food-related risks (University College Dublin).

Notes

1 Interviews with the scientific community were carried out as part of broader fieldwork which involved a range of governance actors. For more details, see Regan (2019).

2 For an in-depth analysis of stakeholders’ perceived risks and benefits of digital agriculture, see Regan (2019).

Appendix

Semi-structured interview schedule

Table

Introduction• Brief overview of the study• Ethical details; information sheet; consent form signing

Conceptualising Smart Farming• Participant’s interpretation of the term ‘Smart Farming’• Present written definition and image: Responses (probe: agreement/disagreement, why)

Exploring Perceived Risks and Benefits of Smart Farming• Main benefits / risks introduced by Smart Farming o Follow-up with scientific community respondents: addressing risks through research (probe: what, how, why)• ‘Winners’ / ‘losers’ in Smart Farming (probe: who, why)

Exploring current alignment to RRI in publicly-funded Smart Farming research (scientific community respondents only)• Views and personal experiences of publicly-funded research and innovation in Smart Farming (probe: funding process, type and composition of projects, translation of findings)• Importance given to anticipating positive/negative/unintended consequences of Smart Farming R&I (probe: why)• Responsibility for anticipating positive/negative/unintended consequences of Smart Farming R&I (probe: who, why)• Processes/mechanisms used to anticipate consequences (probe: personal experiences)• Importance given to concerns/views of end-users/wider society within Smart Farming R&I (probe: why)• Responsibility for engaging non-scientists in Smart Farming R&I (probe: who, why)• Processes/mechanisms used to engage with non-scientists (probe: personal experiences)