On this last day of the e-consultation my task is to summarize the salient points of this discussion and recommend priority aspects for the face to face consultation in July in Bonn. I do not find this very easy, as all participants have brought forwards very interesting and varied ideas, and I cannot possibly do all of them justice. This is why I have to excuse myself in advance: I have not done justice to you all, and this contribution has become much too long. I hope it is nevertheless a worthwhile starting point that allows to harvest your comments, criticism and additions.

The points that I will bring forwards will focus mostly on the discussion that we had in the first two days. I feel less at ease doing practical recommendations which was the topic of the latter two days. It seems to me that we should first  get clear on what the pros and cons of different scenarios are before we can say anything about what we should do. I would therefore like to use the meeting in Bonn to further elaborate different scenarios that came forwards during these e-consultations, and explore their pros and cons together. It would be great to have as an end-product a catalogue of scenarios that describe alternative collaborations around digital farms for the future, which will help innovators to reflect in a broad and varied way about the possibilities this innovation brings. These scenarios should ideally be coupled to suggestions for the further development of legal approaches that suit them and maybe also examples of technological means that support them. (Peter Johnson, Valeria Pesce, Lee Babcock and myself seem to be open to think about technological possibilities as well, but some contributors to the consultation were not enthusiastic about the proposal to use technology to contribute to solving societal problems)

What do you think of this suggestion for our meeting in Bonn?

Scenarios would, in my view, have to focus on (1) how we want to shape (digital) collaboration and (2) how we manage the data. Both are connected.

Open data/data sharing. Data management is an important theme all through the e-consultation. In later days of this e-consultation open access is put forwards as the attractive prospect to pursue, as Nicoline Fourie’s start of day 4 suggests. In earlier days some contributors defend open access too, such as Jacques Drolet, Peter Johnson, Chadwin Reno and Jeremy de Beer. They defend it on the grounds that open data support cooperation and diminish (or discourage) competition between farms, and allow farmers to be more on an equal footing with the large firms.

Francois van Schalkwyk also values open access in different contexts: it fosters transparency and accountability in governance; replicability, verifiability and efficiency in research; and democratised access for social and economic development. But according to him it is not valuable in all situations. Sometimes it is more pragmatic to give open access to some data, but not to all.

But there’s also criticism on the idea of open access. Hugo Besemer, specifies in a contribution on day one what open data are: ‘Open data is data that can be freely used, reused (modified) and redistributed (shared) by anyone’. He argues –based on a Kenyan report- that not every type of data qualifies to be open. For example, data that point out that farmers are financially poor, do not have to be shared with everyone. Ajit Maru, however remarks that it may be important to share this information with a bank when these farmers ask for a loan, but not with all others (such as the milkman). This opens a discussion about what are data that can be open access, and what data should be shared selectively or not at all.

Ben Schaap adds that we need to take into account also whether organizations are still able to function with a specific degree of openness in the data (too much or too little openness may hinder survival or functionality of an organization). Chadwin Reno remarks that making data open costs money, and asks who should pay for services to make data shared or open, Maru and Robert Katende bring forwards that not everyone possesses the capacity to understand the data (farmers, consumers), and Sipiwe Manjengwa, Geoffrey Wandera, Ajit Maru, Foteini Zampati and Francois van Schalkwyk are concerned about the selection of people who have a role in deciding what data are shared and with whom, and note that we should make sure that farmers have a role at the negotiation table as they are often marginalized. Joshua Toews and Andy Dearden, however doubt whether such a negotiation containing farmers will be fair, as not everyone will have the same negotiation power.T

his brief repetition of the discussion formulated on day one, probably is too focused for this summary. However, I wanted to bring back to mind some of the opinions defended in the e-consultation about whether and what data should be open, what data should be shared and with whom. Not everyone agrees that data should be available to all, open access. So, probably this discussion invites a deeper reflection about what the pros and cons of it are, and how we want to collaborate. Sharing data means giving other people the chance to do something with these data.  It is inviting an interaction with these other people. Therefore, it seems to me that it is worthwhile to ask with whom people desire to interact and for what purposes. Is it enough to simply give data to others, regardless of what they do with it? Or does interaction need a basis of trust and –if the answer to that is ‘yes’- what are then the preconditions for trust?

Reflection about whether we want to share data, what data we want to share, and with whom, requires to consider how we want to cooperate. Contributors to this e-consultation seem to start from very different suppositions regarding this cooperation. Communities of cooperation may be large or small, localized or global, rooted in history or a-historic (and a-cultural?).

Some contributors imagine communities with digital technologies to build on their traditional ancestors. Digital technologies are sometimes introduced as if they revolutionize farms: data would offer farmers knowledge that they can act on and make the traditional knowledge that farmers used to share in their communities obsolete. But some authors define agricultural communities as communities that share knowledge. Juanita Chaves, for example, thinks it is important that agricultural communities continue to share knowledge: this can be high tech knowledge, but also traditional knowledge. Chaves as well as Robert Katende see it as an important advantage when communities  share and appreciate knowledge, including also traditional knowledge. A high-tech community would have to preserve this in their view.

The imagined size of communities also differs. Some contributors seem to start thinking about cooperation on the basis of the entire world population such as Ahanda Sosthène Nicaise does who is worried about valuable knowledge and data going from Africa to other parts of the world without African farmers benefitting from it. Equity, fairness, just distribution and inclusiveness is important here, as well as rewarding people for their contribution. Other contributors tend to think about data sharing in a local way, such as Leanne Wiseman (and myself), who states that we should start with the communities that farmers themselves build. Wiseman writes, for example, that ‘(..) the strong relationships that are built by farming communities are a vital first step in the discussion about who is getting the value from the farm data - certainly in many instances we are seeing that third parties are getting the value at the expense of the data contributors who are working on the land.’ 

Furthermore, contributors sometimes contrast collaboration with competition, and some seem to want to  remove competition from the socio-economic system. This is the ‘cooperative mode’ that Peter Johnson mentions, as well as Jacques Drolet. Robert Katende and Uchenna Ugwu, however, write on day 2 that they think there’s no harm in letting companies profit from the advantage they may have from the use of data. They want to balance collaboration with competition. Lee Babcock also gives examples of combinations of combinations of public and private services to give examples of scenarios in which both can interact, serving to realize the public good as well as private ends (profit).

These are examples of ways of thinking about communities –amongst others- that came forwards in the e-consultation. Depending on what the limitations are of the communities that we want to create/foster (and whether these communities have limitations or if it is simply a ‘kingdom of humanity’), we will probably think very differently about what data we are willing to share – and whether we see dangers in the sharing of data. In this respect the theme of trust, also mentioned by the contributors to the e-consultation such as by Robert Katende and Juanita Chaves, is important. The preconditions for trust in data sharing are likely very different depending on the relationships we engage in, inviting also different considerations regarding the development of mores, principles, rules and laws that offer prohibitions and function to define what counts as ‘misuse’ of data.

No law yet. A major gap was introduced right in the beginning of this week by Ajit Maru who pointed out that related to farming and agriculture, there are yet no examples of specific national policies or laws that concern with generation, flow, sharing and use of data. There are only non-binding charters and instruments such as through financial support to share data of and with farmers. Contributors to the debate inquired what would be the right timing to shape such laws: when laws are installed too soon, this would leave little space to experiment with possibilities to collaborate in the network around digital technologies.  I would be in favour of postponing the conversation about laws too, as laws and prohibitions serve to preserve trust in a community, and we first need to explore what communities ICT in farming allows to shape and what is needed to help foster trust in that community. However, the e-consultation offered many interesting threads to think about regulation/law, which brings to mind very different communities.

What plays a role in reflections about rules and law is, for example, the theme of data ownership. There is discussion as to whether it makes sense to speak about data in terms of ‘ownership’. Some participants think it does, such as Serah Odende who remarks that this way of speaking connects well to farmers themselves who look at data collected at their farm as ‘their’ data. This explains also why leaving those data to be interpreted by ICT specialists is such a sensitive matter. Other contributors, propose to explore the meanings of the concept in a more creative way, such as Francois van Schalkwyk who thinks of ownership in terms of labour, implying that stakeholders who do something with data, change something about it and therewith develop ownership rights, or Ahanda Sosthène Nicaise who suggests to look at data in terms of copyright or intellectual property rights.

There are also contributors, however, who do not want to think about data in terms of ownership, such as Jeremy de Beer who propagates open data and tends to think that data should be thought of as a ‘common pool resource’, rather than as individually owned. Similarly, Martin Parr argues –referring to several sources- that there are better ways of tackling misuse of personal insights and privacy than closing down data access due to 'ownership': it is better to install regulations and prohibitions on unwanted activities, than to attempt to restrict the flow of data through assertions of data ownership.

Other themes related to regulation include informed consent. Wiseman argues, for example, that farmers should not be overburdened with the responsibility to gather and appreciate information and then decide to share data. She pleads that ‘the onus needs to be placed on the agribusinesses to reveal the way in which they plan to control, store and manage data collected in simple easy to understand ways, rather than placing the onus on farmers to be expected to understand the intracies of data contracts.’ At the very least, Wiseman thinks  farmers should be entitled to ‘portabillity rights’ in relation to their data: i.e. to be able to have their data returned so that they can make use of it themselves in the future. Furthermore, in a different contribution, Wiseman criticizes the theme of privacy which comes forwards a lot in legal discussions about data: but, privacy law governs personal data or personal/consumer information and Wiseman argues that it is unclear to what extent agricultural data would qualify as personal data.  

Some questions

• Should we strive to realize laws, or postpone it (and first explore possibilities to cooperate)?
• What goals should these laws serve?
• What are valuable/desirable ways to cooperate in the network around smart farms?
• What values are goals for/constituents to this cooperation? What determines the success?
• Should we think globally or locally about cooperation? (When does it make sense to think globally, when should we think locally?)
• Who are the insiders and outsiders of this cooperation?
• Is the cooperation fostered by open access/ data sharing?
• What are preconditions for trust in open access/data sharing in this cooperation?
• What role can laws/regulations play in fostering trust?
• What are pro’s and con’s of different degrees of openness regarding data?
• With whom should data be shared?
• Who should bear the costs of making (some of the) data open – or sharing them?
• What impacts can/will the sharing of data have on the functioning of the organizations involved? And are those changes acceptable?
• Who can participate in the negotiation about what data are to be shared with whom?
• How can the negotiation be fair, given that participants have unequal bargaining power?
• What is a fair distribution of benefits of smart farming, and how can data sharing or access support it?
• What benefits do farmers deserve when they collect data at their farms and share them?
• What are the rights of farmers and other contributors to the network around smart farms with respect to the data?
• Can technology play a role in accounting for the societal needs and values? What roles can it play and what are the limitations of the technological possibilities regarding the societal concerns?

What a complex and large topic we have today! I feel rather speechless and don't envy Leanne Wiseman who started today's e-consultation.  I like a lot in what I read in Wiseman's start and other people's replies, but find it very difficult to react too. I find it difficult, because I don't know what the current scenario is (which scenario? whose? And where?) and what the future scenario should be (Valeria resumed a whole bunch of them very well). So, what is the present starting point and what is the goal? If we don't know that, it's going to be hard to answer today's question....

So, I'm simply going to share how we do it, and I'd love to hear what you think about it or how you do it. Since March I am leader of a workpackage on ethics in a large-scale IoT pilot in Europe (IoF2020) and we chose not to start from the big picture, but to start with smaller ones. Our project includes 19 use cases, divided over different 'branches' of farming: arable, meat, vegetable, fruit and dairy. The ethics part starts with questions that partners in the use cases ask, for example about whether they should share their data with the other partners in the network. Starting from these questions we want to talk about what impacts they want to realize, for their own organizations as well as for society. And then we start thinking about how we get from the present situation to the future one. If the partners in the use case favor certain impacts, this imposes requirements on their cooperation, and on the technology itself.

So that's how we do it. We start in use cases and from there we try to acquire insight in the values at stake and the impacts participants want to realize. And then we think about what values the technology should serve (value-sensitive design is the approach) and how the collaboration between the partners around it should be organized.

From these individual use cases we can generalize. Hopefully at the end of our project we can offer a catalogue of designs for IOT that serves different collaborations, so that stakeholders have something to choose. And guidelines and codes of conduct should be assessed to see whether they support the desired collaborations.

This approach has downsides, of course, but what I like is that it allows for different solutions in different contexts. And it starts with needs and values of people in the network around smart farms, as well as farmers themselves. From there we generalize towards designs of IOT that can be implemented in different locations, and guidleines and codes of conduct that help future innovators to think about what they want in their contexts. 

Thanks to all participants for the valuable contributions. I just have a few things to add.

1. With respect to desirable scenarios. Data-driven agriculture can take shape in many ways, and probably there's not one desirable scenario for all contexts, but we should imagine various scenarios. Depending on the needs and values that play a role in a specific contexts, it must be possible to prefer one or the other. (In Europe, for example, farmers may be reluctant to share data, because they are afraid that their competitors will use them; but maybe in other contexts it makes farmers stronger if they share data? I lack knowledge here, but would love to know. Maybe some of you can fill me in here...) 

2. Regarding the concept of 'success', which I find very interesting, I here list impacts ascribed to data-driven farming that I encountered in the literature on ethics of smart farming and that may be good candidates to think about success. Success can  be interpreted in a societal way: it can mean, for example, that farming imposes less of a burden on the environment, that animal welfare at farms is improved, that the production rate of farms goes up and farms will be able to feed more people, that the acceptance of farm products by consumers is improved, that consumers make better food choices (healthier, more environmentally friendly, better for animals), that the autonomy of consumers is fostered, that the chain leading from producer to consumer becomes shorter (for ex. localized in one region),  that food security is fostered, that the profession of farmers becomes attractive again to young people, that empty regions will have more farms and be populated again, that there will be more jobs. But success can also be interpreted from the perspective of individual farms: success can mean surviving as a business, remaining competitive, staying autonomous in the decisions taken at farms, continuing to have satisfying work, having less administrative proving that you live up to the  environmental rules, that there is equal access to technologies needed to realize data-driven agriculture, and maybe more....?  

So, it seems to me that if we want to talk about desirable future scenarios, maybe we should get clearer on the position from which we start to think about it (our background, culture and values, needs) as well as the goals we seek to realize, and the perspective from which we look at the goals (from a societal perspective or as individual farmers). It may be that from a societal perspective we are interested in different goals, than from the perspective of individual farms, although it is also possible that both overlap. Open access may be attractive, for example, to look at whether and to what extent farms contribute to the societal goals (less burden on the environment, more production for more mouths to feed), while individual farmers may not be so keen to share those data with everyone as it may kill their competitive position in the market.

So, what is success? There is, I think, a lot of conflict in our ideas of success, which may support differing conclusions as to how data need to be shared and with whom. It seems to me we need to get clearer on those differences, and maybe create various scenarios that suit them, in order to facilitate shaping innovation strategies in a more informed and reflected way....?

What challenges are related to accessing data and sharing data? In his opening to this e-consultation, Ajit Maru brought forwards important challenges that are often mentioned in connection to data access/sharing: that is, lacking laws and policies regarding how we need to deal with farm data. The issue of 'privacy' often comes up in relation to data, and Maru brings it forwards too, in relation to the recent discussion about Facebook-data.

As my expertise lies in ethics, I would like to elaborate on this a little further. It seems to me that we should not set the laws too soon, but first leave some time for small-scale experimentation. If we formulate laws that protect the privacy of, for example, farmers, then a lot of the possibilities that ICT offers to shape new collaborations between stakeholders in a smart farming network will no longer be possible.

In the literature on ethics of smart farming -which is at present still limited- privacy issues are often brought forwards by authors who talk about data in terms of 'ownership'.  Who owns the data? That is the main question that they ask about data. Ownership can be looked at in different ways. Farmers frequently think that they own the data, as they collect them at their farms. Let's call these data the 'primary data'.  Yet, the data are not worth much if they are not interpreted by ICT specialists who develop the algorithms that establish menaingful connections between large amounts of data collected at different farms. The interpreters of the data often consider themselves the owners of the 'computed data'. Ownership comes with rights: the owners of something can use it for their own benefit. Related to ownership questions is therefore the question who can use data for what purposes? Can ICT specialists use the computed data to inform their decisions on the stock market, for example? And how can farmers be ensured that their data won't be shared with their competitors?

Understanding the challenges in terms of 'ownership' issues, calls for protective measures. Privacy-laws protect the rights of the owners. Privacy issues come up here as property is often considered private and owners can decide about their property as they see fit.

But talking about data in terms of 'ownership' also limits the possible ways in which parters in a  network can fruitfully cooperate. It is also possible to approach farms as actors that have a role in realizeing societal goals. This would lead to a different way to look at issues related to data access and sharing, such as for example: who should have access to data if we want to realize the societal goals of smart farming? In a European context that I know best, these public goals can be -for example- to check whether farms diminish their impact on the environment, to enhance public acceptance of production processes, or to to foster food safety. Considering these issues it may be worthwhile to give governmental organizations access to data, or consumers, for ICT could foster new -less labor intensive- ways to monitor the environmental friendliness of farms or to establish fruitful exchanges between producers and consumers about -for example- animal welfare.

It is important to set laws, such as Maru proposes. But in my view we should not do it too soon, as this risks to close off possibilities that smart farming offers. The 'smartening' of farms also opens up a possibility to reconsider the components of the social networks  around farms, and the types of relationships between them that we want to foster. It would seem wise, therefore, to carefully consider the pro's and con's of different ways to shape these networks in varied contexts, before we create the laws. In this way laws are able to support the relationships that we want to bring about around smart farms, and they will not impose obstacles to them.

Issues relating to data access and data sharing would therefore have to focus first on the type of relationships we want to establish, and why. After that we can consider in a more informed way how we can prevent unwanted use (or misuse) of data and seek ways to prevent that, such as by means of laws.