Fair trade in building digital knowledge repositories: the knowledge economy as if researchers mattered

This Special Issue addresses the question of whether we can trust Digital Knowledge Repositories (henceforth DKRs) in the field of biomedical sciences. An interesting fact about these repositories is that many are envisaged, and many projects launched, but few manage to achieve the volume and uptake needed for taking off as useful research and translational tools or to keep up with the new knowledge constantly published (Baumgartner et al. 2007; Tripathi et al. 2016). So, other questions arise: can we trust that promoters of DKRs will manage to realise their good ideas? Can we trust that a community of experts has the means and motivations for sustaining the growth and continuing update of a new DKR? The risk of failure is high. A lot of attention has been devoted to the risk that users may or will not trust the DKRs, and this epistemic trust is surely an important issue. But there is another risk that has been much more neglected: that the knowledge producers may not trust that the effort to contribute their knowledge to the DKR is worth their while. I propose the concept of expediential trust to identify and draw attention to this phenomenon.

My argument is fairly simple. It starts with the observation that many DKR initiatives are launched relying on unrealistic expectations, because of a wrong assumption about researchers’ motivations and freedom in managing their time. The result is that many attractive projects struggle to be realised and have modest success or fail.

Both existing literature and the case study presented suggest that even very valuable and well-thought projects struggle to win the support of the providers of the knowledge to be collected in the repositories. Indeed, the case study shows that even when the project has the qualities needed to win contributors’ support and shows concern for them, their willingness still cannot be taken for granted.

I argue that the problem of expediential trust is not likely to go away, instead it is actually deeply rooted in the changed circumstances and working environment within which researchers, and especially academic researchers, operate. The entrepreneurial nature of universities and the economic mission of science and research have generated heightened competition, together with a pressing need of a very strategic management of time and objectives by researchers. Endeavours based on altruistic motives or on distant and uncertain rewards can only be supported at a cost and even at a disadvantage. Furthermore, the academic workforce is stressed and strained: growing responsibilities and demands together with increasing job insecurity have skewed the work-life balance. Incentives for the contributors need to be understood in light of the constraints and counterincentives at work in these circumstances, and with a sympathetic eye for the declining well-being of researchers.

Very often the promoters ofDKRs are driven by a compelling view of the possibilities that would be opened up by collecting, structuring, organising, comparing, computing and making easily accessible large amounts of data. This effort is imagined to be able to empower researchers to do things hitherto not possible or extremely time consuming and labour-intensive. New possibilities would open up, more efficiency and speed will be within reach. So, the DKR looks to them as an attractive ideal for a community of users, and since usually among the users there will be those very researchers that are now expected to share their knowledge or data through the repository, the assumption that they will have an effective motivation to do so may seem almost obvious.¹ After all, bringing about the DKR is in the contributors’ own long-term interest, and as scientists they are believed to be rational people. Furthermore, their contributing is consistent with the practice of sharing that is part of the scientific ethos. Logical as it may seem, the reasoning behind this motivational assumption has often failed to stand up to the test of experience.

Bruno Strasser (2017) has studied some of the earlier and higher profile repositories developed in the domain of the biological and biomedical sciences, like the Protein Data Bank and the GenBank, and has found that the attractiveness of the ideal is not sufficient to motivate researchers to contribute their data. He explains the reluctance invoking the ethos of experimental scientists, who “valued individual achievement over collective participation”, so that it is particularly challenging to persuade these researchers to share “data difficult to produce and potentially rich in the epistemic rewards of new publications” (Strasser 2017, p. 190). As a result, in the cases he studied, the rate of contributions from scientists was much lower and slower than the promoter had expected and hoped. In order to boost participation and contribution in the endeavour it was necessary to acknowledge the individual need of researchers to achieve recognition for their work through peer-reviewed publication. So, in order to leverage their contributions, the promoters of the repositories had to win the support of the scientific journal editors, who eventually agreed to impose as a condition for accepting submissions that the authors had submitted the data to the relevant repository. Thus, the individual reward had been made conditional to the act of sharing. Therefore, for Strasser, it was not the vision of a collective benefit or an ethos of sharing that prompted the contribution of their data, instead “researchers have shared data because it became in their own interest to do so, as defined by the existing reward system in the experimental sciences” (Strasser 2017, p. 198). And that reward system as well as the ethos of experimental scientists, Strasser claims, is not communitarian, but individualistic.

What is important for our argument is that Strasser’s case studies show that spontaneous and disinterested contribution was not forthcoming because it was not in line with the existing incentive structure. In order to boost contribution, a targeted alteration of the reward system was needed. Several other studies have confirmed that attempts to promote data sharing continue to meet with limited endorsement from researchers (see Enke et al. 2012; Tenopir et al. 2011 for the results of large surveys). For instance, Hedstrom and Niu (2008), Costello (2009) and Borgman (2012) attempt to understand why in the face of many compelling reasons for data sharing, researchers still resist endorsing sharing practices. All these authors list some of the most important reasons that keep researchers from sharing data, and discuss the problem of lack of incentives and the misalignment between the interests of data users and data producers. However, as it will become clear from our argument, we find more appealing Hedstrom & Niu and Borgman’s approach that stresses the importance of having policies that benefit data producers, rather than Costello’s approach that is oriented to impose data sharing requirements (but he also emphasises the importance of providing resources) and questioning the ethos of non-sharing researchers—it is worth noticing that Hedstrom & Niu show that compulsive measures have limited effectiveness and are poorly enforced. Another striking example of the resistance to data sharing is offered by Savage and Vickers (2009) who performed a little experiment and tried to get the data from the authors of papers published in journals that requires data sharing (PLoS Medicine or PLoS Clinical Trials). They tried to contact 10 authors and from the 8 that they could reach, they manage to get the data only from 1.

The importance of the regime of incentives and rewards within which expected contributors work emerged also in the case study that I present in the next two sections below. The significance of this case study is that the project behind the DKR looked like the kind of initiative especially likely to—and quite considerate in the attempt to—win contributors’ support. Yet, in spite of all its qualities, it still run into the challenge of eliciting sufficient contributions and I conclude that uncertainties around researchers’ willingness to contribute is the major challenge to its success.

The persistent difficulty that DKR promoters encounter in recruiting contributors persuaded me that new perspectives and ideas are needed. At the same time my experience in academia—where I have witnessed how heavy workloads and career demands can be—suggested to me that responsiveness to contributors needs and well-being is both ethically required and instrumental for DKRs success. So, I propose a knowledge producer centred approach.

Rather than sticking to Strasser’s hypothesis about the individualistic ethos of the experimental scientist, I focus on the structural changes in the institutions, drivers and political economy of knowledge production, and on the impact that they have had on the life and wellbeing of researchers. Building on an extensive corpus of research on the changing nature of scientific research and institutions, I formulate the hypothesis that such profound changes have introduced tight constraints to what researchers can do without suffering a competitive disadvantage or taking up burdensome and hardly sustainable additional responsibilities. I do not take any moralistic attitude towards researchers and their motivation, on the contrary, building on a growing body of evidence of the strains and stress suffered by academics, I take a sympathetic look at the pressures and demands under which they currently work. Taking their struggles and needs seriously, I advocate the importance of fair working practices (or production processes) around knowledge sharing—a kind of fair trade approach for workers in the knowledge economy.

At the level of individual motivation, it is a well-established principle of many coaching and counselling techniques to recommend to people to get a taste for the process that leads to a result, rather than trying to be motivated by the lure of the distant goal in order to endure current and forthcoming sacrifices. The distant goal is too remote a reward, hence, to sustain an effort, some current benefit is needed. If we adapt this model to data sharing, we need to make some adjustment and not simply expect that researchers will come to like the work required by data sharing—as Hedstrom and Niu (2008) observe, “it is hard to imagine how to make it fun to create ‘archive ready’ data”—but rather take steps to make sure that they will receive some external benefit (extrinsic rewards) from incorporating that work in their practices. With this kind of adjustment, I believe that this approach should be applied to the production of DKRs. Namely, promoters of the idea should abandon the final endpoint attraction strategy (the gravitational pull model, so to speak), and instead adopt a producer-centred perspective, and a generous one, inspired by the ideals of fair trade: data are a product that should be valued and rewarded adequately and some provision to improve the working life of the producers should be put in place. The knowledge economy has been touted as something good because it would generate highly qualified and attractive jobs, but, as I show in Sect. 7 below, the reality is that the quality of life and even the mental health of knowledge workers is steadily deteriorating. Increased demand for knowledge has not brought benefits to the workers supplying that knowledge, but only added weight to their workload. With this concern in mind, I propose the concept of Expediential Trust as both a useful analytical and practical tool. Expediential trust helps in understanding not how to compel people to contribute, but how to make it safe and rewarding for researchers to contribute. So, the concept of expediential trust is proposed as a tool, a fictional device to help in setting up a fair and sustainable process of knowledge sharing.

The concept of expediential trust is meant to describe the kind of trust that is needed to sustain a commitment and effort from people who have to manage their time and workload in light of strict targets, accountability to their employers and tight career management demands. It is proposed as complementary to the concept of epistemic trust. While this latter focuses on the epistemic requirements that are needed to win the approval and trust of scientists and other professional experts, the idea of expediential trust points to the incentives and bonuses that experts and scientists need to have in order to endorse a project that makes new demands on their time and efforts. Let me try to make this distinction as clear as possible. Kankanhalli and colleagues have noted that.

Success of EKRs [Electronic Knowledge Repositories] requires that knowledge contributors be willing to part from their knowledge and knowledge seekers be willing to reuse the codified knowledge (Kankanhalli et al. 2005, p. 115).

Epistemic trust looks at the conditions needed for knowledge seekers to be willing to use the knowledge offered, while expediential trust looks at the conditions needed for knowledge contributors to be willing to make their knowledge digitally shareable.

But what kind of trust is expediential trust? It is clearly not a default and pre-reflective disposition, as the trust described in many accounts of social or interpersonal trust (e.g. Luhmann 2017), but rather an explicit type of trust, emerging from reflective awareness of facing a choice. In this respect, it is closer to the philosophical accounts of trust that ask whether trust is well-founded, but rather than focusing on the cognitive and epistemic element, it revolves around the expected payoffs. So, it is triggered not by an epistemic doubt: “shall I trust this belief or report?”, but by an expediency doubt towards a collective endeavour: “shall I trust that joining a collective effort will pay off?” (an analogy can be provided by the choice of a pension scheme or health insurance, or whether to join a housing cooperative). It is a question that arises when there is a time gap between a demand for a contribution now, and the expectation for a reward later, provided that other people too have contributed to the collaborative scheme and made it viable and sustainable. Contrary to some basic social practices, where the cost of opting out is so great that people trust by default—because it is too costly to distrust—the question of expediential trust emerges when not joining a collective endeavour is clearly a practicable option, hence joining involves a risk and reasons to take it. This kind of trust can be analytically divided into a technical and an ethical component. We may question whether the collaborative scheme is well designed and suitable to deliver the expected result, and we may also question whether the other participants in the collaborative scheme will support it consistently and dependably. Furthermore, there is an additional ethical dimension: is the scheme distributing burdens, risks and rewards fairly? This is the central concern at the basis of the idea of expediential trust. By postulating that people can safely keep out of the scheme, expediential trust invites us to frame the choice so that it has to be attractive and fair for people who join the collaborative effort. People should join because they trust that the collaboration is designed in such a way that it makes their condition better than their current one, instead of joining because they are afraid of being left behind or penalised if they do not join, which is the mechanism used by the compulsory policies.

However, the most important thing is that the concept is not proposed as a heuristic tool to be used by the people who are facing the actual choice, namely the researchers. Rather, it is proposed as a kind of thought experiment, as an intellectual device to help the imagination of those who are shaping the practice: how do we need to design the practice, its rules, demands and rewards so that people evaluating the practice will be willing to join and sustain it? Given that John Rawls’ fiction of the original position (Rawls 1999) is so well known, it can be used to illustrate, by analogy, the role that I attribute to the concept of expediential trust. Rawls proposes the idea of the original position as a fictional choice, as a device of representation, i.e. as a test for assessing the fairness of principles of justice for the organisation of society. If we want to work out fair principles for society, we need to design them so that they will be chosen by hypothetical citizens who are self-interested but that they do not know their position in society. Rawls has very often been misunderstood as claiming that real people are like people in the original position, but this is a misreading. Expediential trust is proposed to perform an analogous function: it is meant to help to design collaborative schemes for knowledge sharing that hypothetical researchers free to choose whether to join or not would find appealing and rewarding. So, the concept is addressed not to real researchers (and neither assumes that they are self-interested nor recommends that they should be), but to those designing the schemes for knowledge sharing, with their demands and incentives. Real researchers are hopefully motivated by other expectations, for instance the desire to participate in an epistemic community and support it.

The concept of expediential trust is meant to help those shaping science policy and knowledge-sharing projects to design their plans as if they could win expediential trust from researchers. By doing so, we expect them to achieve two results. First, they can better achieve their goal by winning researchers’ compliance. Second, they can achieve their goal not through imposing additional burdens to the producers of knowledge, but by making their working life better. Thinking in terms of expediential trust can make knowledge-sharing policies and projects both more successful and more ethical. To wit and paraphrase the subtitle of Schumacher’s (1973) famous book (Small is beautiful. Economics as if people mattered), the motto behind expediential trust and fair knowledge sharing would be “the knowledge economy as if researchers mattered”.

Raising the question of expediential trust does not mean claiming that researchers are only moved by self-interest, that they are textbook examples of homo economicus. Nothing could be further from my intentions. Rather, the question of expediential trust leads us to ask whether researchers can afford to take up new and demanding commitments when their duties are tightly specified and when they are at risk of being out of the research game if they do not manage their priorities and efforts strategically and efficiently. The successful realisation of DKRs needs to be, and to be trustfully perceived to be, in the interest of those researchers at the coalface, or, if you prefer, at the shopfloor level. Just like administrative reforms will not work if they do not win the support of street level bureaucrats, lofty scientific ideals will not be achieved unless researchers believe that it is wise and sustainable for them to put work and time in their pursuit. The transformation of research management forces DKRs promoters to think in terms of expediential trust, lest they pursue illusory and futile goals. Furthermore, the deterioration of the wellbeing of academic researchers prompts DKRs promoters to think in terms of expediential trust, lest they want to achieve their goal through making more miserable the life of their knowledge providers.

Can we instead pursue better science and increased societal benefits while making the life of knowledge producers better rather than more difficult? Our point is simply that techno-moral change (Swierstra 2013)—i.e. change in practices and moral judgments that are partly triggered by adopting new technologies—that looks desirable comes with some costs. Ultimately these costs are borne by people, and this raises two important questions: are these costs too high and imposing unjustifiable misery on some people? Are the costs distributed fairly?

In this section I argue that the current institutional circumstances within which research is carried out generate a structural misalignment between the expected sharing ethos of scientists and the structure of incentives under which they work. A consequence of this misalignment is that knowledge sharing may be promoted counting on an incomplete and partially outdated understanding of the ethos of researchers and of their current working circumstances. This could result in policies and strategies that disregard the needs and constraints under which researchers work, impose new burdens on them and even point a moralising finger at them.³ To prevent these outcomes, I have proposed the concept of expediential trust as a thinking device for imagining the conditions under which researchers’ collaboration would be voluntary and fair.

The transformation of science and research have been widely studied and discussed and one thing about which there is very substantial consensus is that science and research have been practiced under rather different institutional circumstances in the early decades after the war and in the last decades. Here I focus on universities in particular, although of course not all researchers work in academia. However, academic researchers represent a very substantial group, and some of the transformations described (like increased competition and strict accountability regimes) apply widely across the research institutions. What changed most obviously were the arrangement and management of the research organisations: their mission, funding and accountability; as a result, the practices, the ethos, the incentives and career structure profoundly changed as well. John Ziman in his Book Prometheus Bound: Science in a Dynamic Steady State (Ziman 1994) tried to convey the extent of these changes through the thought experiment of imagining a researcher coming back to her department after 30 years of space–time travel and notes how she would be exposed to a number of new concerns and worries that are now prominent in the mind and talks of her colleagues: from the attempt to align with R&D policies and priorities to how to compete for scarce resources, from concerns about performance indicators to the need of managing a portfolio of skills (teaching, research, outreach, administrative, fundraising, collaborative, leadership are commonly expected skills from applicants for academic jobs or research funds) in order to secure the next, non-permanent job and so on and so forth. Ziman concludes that.

In less than a generation we have witnessed a radical, irreversible, world-wide transformation in the way that science is organized and performed (Ziman 1994, p. 7).

Furthermore, he stresses how these changes have penetrated deeply into the daily practices and the culture of research, to the effect that they are now part of “the climate in which scientists have to live” (Ziman 1994, p. 5). For our current purposes we do not need to be very precise about when the transition took place—most likely these changes started in the late Sixties, accelerated in the Seventies and Eighties and were globally well established by the mid 1990s.⁴

What is important for my argument is that we now live at a time when the values and culture of research suffer from a strange form of dissociation between the lived experience of researchers caught between the pressure (and necessity) to achieve their contractual goals (deliverables), boost their scientific production, expand their network, carefully plan and manage their career and daily schedule on one hand, and on the other hand a dominant rhetoric still powerfully influenced by the scientific culture and values that prevailed before the recent institutional transformation. This is not only because often cultural change is slow and there are survivals, but also because in many ways those values are still compelling and attractive. I see colleagues (and myself) motivated by and committed to these values—although these motivations and commitments are bounded by a grid of external constraints and new interiorised goals and values. But these values and ideals are no longer the only and dominant defining values of researchers. Researchers are often experiencing a painful value schizophrenia: to put it bluntly they have both a scholarly, idealistic-perfectionist ideal and an entrepreneurial, efficiency-productivity ideal. They are moved by curiosity and by the desire of deploying all their skill and ingenuity in their scientific activity and production, by the desire to do it to perfection; but, at the same time, they want to incarnate and express those broader skills that make for a successful researcher. They want this not only because of ambition, but because the profile of a successful researcher under current circumstances is a fascinating ideal too. It requires a very complex set of skills and it can appear more complex, balanced and less obsessive than the purely scholarly ideal.

So, here we are seeing neither a conflict between “pure” scholarly values and “corrupted” scholarly values, nor a conflict between the smugness of ivory tower scholars and the pragmatic scientist engaged in the real world. Both ideals are respectable at their best and subject to excess or degeneration. What is important is that their coexistence creates tension and sometimes irreconcilable demands. Sometimes a satisfying and sustainable compromise can be found, sometimes one ideal needs to be—temporarily, it is hoped—forfeited.

However, it has to be acknowledged that the older set of scholarly values are often perceived as loftier, nobler and purer. Furthermore, they are associated with the dominant scientific ethos of the times when science established itself as a driving force of modern society and as the most respected source of authoritative knowledge. Arguably, they are also needed to support the prestige and credibility of science. The noble ideals and the virtues needed to realise them create an ethical image of science practitioners and hence of the practice of science itself. Conversely, the new circumstances of research do not seem to have been accompanied by an ability to articulate a new scientific ethos that can be seen as equally noble, compelling and distinctive. Moreover, the perception that traditional values are still obtaining, represents an asset also for achieving the new mission of research.

However, while in practice researchers have to find an integration, or at least an accommodation, between the old aspiration and the new demands for efficiency and usefulness, at a theoretical level we still have to see a new formulation of the new key values and norms of science that could match the attractiveness and respectability of, say, the Mertonian account.

But what are these institutional transformations we are talking about? The transformations most relevant to our argument are the following: the belief that research plays a vital role in sustaining the success and competitive edge of a country and of its industrial and economic productivity (Etzkowitz et al. 1998; Gibbons et al. 1994; Rossi 2009; Zawdie 2010)—this is often described as the “third mission” (after teaching and research) of universities; from this follows the belief that policy should support research to the extent that research reorients its mission towards economic development and social impact (Guston 2000; Guston and Keniston 1994; Marburger 2015; Ziman 1994)⁵; another important point is that resources allocated to science and research cannot be assumed to be growing and during hard times are subject to cuts, so research is now operating under circumstances of resource scarcity and competition (Sarewitz 1996; Ziman 1994); taken together the previous points lead to the further corollary that the allocation of resources for research will be largely based on research projects’ or research groups’ ability to provide evidence of their impact on economic performance or their contribution to social goals—like security, public health, natural resource management—and to enhancing citizens skills and lifelong learning, cultural development and knowledge-based development (OECD 2007). Finally, these changes have forced universities and other research institutions to improve their efficiency and competitiveness through adopting managerial techniques coming from business, so much so that the concept of “entrepreneurial university” has become quite popular (Etzkowitz et al. 2000; Foss and Gibson 2015; Gibb and Hannon 2006; Guerrero and Urbano 2012; Meissner et al. 2018). University are, as an OECD report puts it, “in the knowledge business, which is a competitive and fast-moving sector” (OECD 2007, p. 13). An implication of this new mission and organisational model is that they need to be “entrepreneurial at all university levels” (Guerrero and Urbano 2012, p. 54), and this of course effects a transformation of the organisation’s culture, of its ethos. It also implies that efficiency becomes a key organisational criterion: to be drivers of innovation and growth universities need to be on top of the game and therefore they need to use their resources, including their human resources as efficiently as possible. A further corollary is a culture of performance indicators, which is of course applied to the performance of teachers and researchers. Needless to say, not all researchers operate in universities, but these entrepreneurial and efficiency demands apply as well in industry and in many—although not in all—other research institutions.

This description should provide a context that helps in explaining why researchers may find it hard to contribute to knowledge repositories if their working contribution is expected only on the basis of the traditional scientific value of knowledge sharing (what Merton (1973) called “communism” and other authors following him renamed “communalism”). Many researchers do see the appeal of that value and are inclined to respond to it, but they may not do so because they are constrained by tight demands and pressures stemming from the insecurity of their position or funding. When their time is carefully managed in order to perform their obligations, adding another time-consuming service may simply be too much: something that they cannot sustainably squeeze in. Similarly, when competition for jobs and funding is very intense and both resources are scarce (as Nikunen (2012) effectively puts it: “The number and supply of researchers is high, while secure positions are scarce”), to devote time and energy to something that your colleagues/competitors are not obligated to undertake may put you at risk of suffering a competitive disadvantage that can drive you out of the game. I thus suggest that both the success of knowledge sharing initiatives and concerns about the fairness of the distribution of the burdens necessary to sustain it demand that the question of expediential trust is heeded and that strategies for generating it are developed.

An academic career was once viewed as offering low stress, secure, safe employment, and high social standing with opportunities to do satisfying, autonomous work (…). Over the past 20 years, the academic environment and perceptions about academic careers, have changed drastically (Catano et al. 2010).

The transformation described in the previous section has not only limited researchers’ discretion in using their time and their liberty to devote themselves to disinterested and altruistic pursuits, it has also profoundly affected their working conditions, well-being and mental health. A growing body of research and of public personal testimonies (Dunn 2013; “Quit Lit,” 2019) show very clearly how serious these trends are. The managerialisation described above—with its increased competition and drive towards efficiency and performance assessment—has contributed to well-documented rising levels of stress and anxiety among academics (Abouserie 1996; Berg, Huijbens, and Larsen 2016; Biron, Brun, and Ivers 2008; Gill 2010; Gillespie et al. 2001; Kinman 2001; Tytherleigh et al. 2005; Winefield et al. 2003, 2008), to declining levels of job satisfaction and inability to achieve a satisfying work-life or work-family balance (Bothwell 2018; Cownie 2004; Kinman and Jones 2008; Pillay and Abhayawansa 2014; Slišković and Seršić 2011), to overwork and burnout (Hogan, Hogan, and Hodgins 2016; Watts and Robertson 2011), to widespread feelings of inadequacy and fear about the future (Ball 2015; Gill 2010; Morrish 2019), to rising levels of mental health problems (Gorczynski 2018; Kataoka et al. 2014; Kinman 2001; Mullings, Peake, and Parizeau 2016; Pace et al. 2019). These results consistently emerge from research carried out since the 1990s in the UK and Europe, North America, Australia and New Zealand. The most frequently mentioned causes for these alarming outcomes are work overload, excessive administrative tasks, insufficient resources, job insecurity, intense competition, and insufficient recognition and rewards.

Once this picture is used as a background for understanding the deep causes of the limited endorsement of the sharing practices required by contributing to the DKRs, things begin to make sense. Researchers feel that they have work overload and that they are struggling to cope with it, and the main obstacle that they report to contributing to DKRs is that it is time demanding (Tenopir et al. 2011). This cannot be surprising. Researchers report distress and anxiety related to competition and job insecurity and another main reason that they report for not contributing to DKRs is the perceived necessity to prioritise their publications and making the most of their data (Borgman 2010; Hedstrom and Niu 2008). Again, this seems perfectly logical. Just as these causal links are obvious, I believe that it is evident that any strategy to increase participation in the sharing practices required by DKRs based only on compulsion, shaming or hypothetical future reputational gains are not only inadequate but insensitive, if not callous and exploitative. Instead, I propose to echo the suggestion made by Pace and colleagues to improve the quality of academic work and argue that efforts to improve sharing “that simply ask academics to do more without reference to their primary work motives” and duties or to the reality of their daily practices and struggles are bound to fail. Only a fair allocation of workload will enable researchers to work better and reduce their psychological distress (Pace et al. 2019). Furthermore, I endorse Morrish’s admonition that “the simple, humane values of kindness and care for each other must be recovered if we are to ameliorate the toxic university” (Morrish 2019).

An idea that was suggested by the promoters of the AOP-KB was that of buying-out the researchers’ time needed to share their knowledge and data. I believe that this points in the right direction, because it represents a genuine departure from the strategies leveraging on compulsion and reputation. Buying researchers’ time is a real and immediate reward, that shows an appropriate valuation and appreciation of researchers’ time and effort. This is in stark contrast with the reputational rewards, which, under the current conditions described above, are reinforcing a logic of relentless competition that makes people overworked, stressed and miserable.

If organisations really believe in the value of data and knowledge, then they should commit the resources needed to produce them, without expecting that they are simply brought about through the motivations of an economy of insecurity and hope. What the reputation-centred strategy fails to appreciate is that in a intensively competitive environment introducing a further scientific productivity metric for building a researcher’s reputation is imposing a burden on researchers, while it is not giving them a direct reward, but only a competitive edge that will disappear once the practice has become a baseline professional requirement. At that point there will be no competitive advantage in contributing, but on the other hand the task will have become an additional standard duty, without any certainty that the additional time and effort that it requires will have been recognised as part of the workload of every individual researcher. To be sure, at this point researchers will have access to powerful instruments that will make their work more productive and possibly also more rewarding, but this increase in productivity will benefit many other stakeholders, which should bear some of the cost of achieving this productivity gain.

Suppose that in some manufacturing industry workers were told that they should work five extra hours per week without retribution to contribute to the optimisation of the production line. Once this will be achieved, there will be a substantial gain in productivity—which will benefit the industry and its customers—and the workers will have the satisfaction of seeing that they will achieve more during their working hours (but they will have to keep working additional hours to keep updated the new system). Who would expect the workers and their trade unions to think that this is a good deal and they will benefit from it? We would surely expect that they would accept only under one of the following scenarios: (a) they get paid at a good rate for the extra hours; (b) they do perform the new task during their normal working hours, not as extra hours; (c) they are threatened to lose their job if they do not accept the deal: the company is ready to relocate into a lower income area where workers will accept to work the additional hours for the same salary. Now, unfortunately leveraging only on the reputational incentive bear no resemblance with options a) and b), but only with option c): because if researchers do not accept, they will suffer the competition from researchers who are prepared to do some extra and unrewarded work. Proposing such a deal to the workers surely does not look like kindness and care.

The digital revolution has also shown us the dangers of relying on voluntary contributions and reputational hopes: it is a way of mobilising many people and pushing them to contribute and to share in the expectation that this will earn them a reputation that in the future will pay off. In this way, their free contribution crowds out the work of paid professional who were providing the same services for a present and tangible reward. Actual rewards are replaced by virtual and hypothetical rewards. Professional performances provided for a fee are replaced by tasks carried out in pursuit of possible future advantages. In the case of DKRs, since they are relatively new kinds of goods, they could not rely on an existing dedicated profession devoted to them, so there is not the crowding out effect: we do not have professionals being replaced by amateurs hoping to become free-lance content providers. Yet, we still have unpaid (highly qualified) service provision expected instead of paid-for service provision—and that is why buying researchers’ time for doing that would be a paradigm shift. On the contrary, the reputational reward feeds the toxic trend of expecting academics to continually expand their skill set and their achievement record at an unsustainable pace. As Berg and colleagues note,

academics must now constantly seek ways to increase their future value, through, for example, successful grant applications, peer reviewed publications (in journals with the “right” impact factor), website blog posts, hits on their personalized socio-scholarly media websites (…), paper citations (…), and various other “measures of esteem” (Berg, Huijbens, and Larsen 2016, p. 178).

This is precisely one of the key drivers of researchers’ stress, anxiety and mental health problems. Can we really blame those among us who resist these pressures to do additional work for free, to give our scarce and precious time, to “fabricate ourselves in ever lengthier and more sophisticated CVs” (Ball 2015, p. 259), who answer, like Melville’s Bartleby: “I would prefer not to”?

There is a puzzle around DKRs: everybody seems to agree that they are very attractive resources and that it would be good to build them, yet a lot of the attempts to realize them lose steam and fail. Like every human and collective endeavour, they can fail for a number of reasons, ranging from lack of resources, poor planning, bad luck or timing. From the perspective of a philosophical analysis of disruptive digital technologies, attention has mainly focused on epistemological issues—and more specifically on epistemic trust and on the transformations brought about by aggregating and standardising knowledge through the support of digital technologies—and the new social network that they create—epistemic communities. Our experience with the case of a quite high-profile digital repository in the field of toxicology, the Adverse Outcome Pathways Knowledge Base, shows that indeed epistemic trust is of crucial importance and a necessary condition for the success of these kinds of projects. However, interviews and engagement with the AOP community revealed that another serious challenge could prevent the success of the initiative: the struggle to attract enough contributors and to keep them committed to supporting the growth of the Knowledge Base. This challenge led me to articulate the notion of expediential trust to help focusing attention on the needs of the researchers who feed knowledge into the repositories. The work needed to support the creation of knowledge repositories is very often voluntary work, done out of an altruistic commitment to create a resource valuable for the research community to which researchers belong or aspire to join. While I do not believe that this motivation has lost its appeal, I think that it may be too remote when challenged by powerful competing demands and counterincentives. Thus, expediential trust attempts to use instrumental rationality to pursue communicative rationality or reasonable fairness—under the guise of conditions of work that would be acceptable for unconstrained rational agents.

From the perspective of the researchers who are supposed to contribute content and curation services to the DKRs, this additional task has to be compatible and sustainable from two different perspectives: (1) from a personal ecology of practices, and from (2) an interpersonal reality of competition for scarce resources. So, before committing to contributing time, work and research findings to a knowledge repository, researchers are bound to ask themselves: is this further commitment to do extra work and share my research results something that is sustainable? Is it contributing to my successfully performing my professional targets and goals? Is it a practice that can be sustainably integrated into my working routine or is it further eroding my already limited workfree time? Is devoting time to the task and sharing my knowledge going to be a competitive disadvantage in my quest for jobs, progression, funding? These, we believe, are questions that is perfectly reasonable for them to ask in the current research and employment conditions. But acknowledging that these are sensible questions for them to ask means that promoters of DKRs need to consider them too, because the support of knowledge producers is necessary for the successful development of DKRs. The promoters are therefore facing an alternative: they can either look only at the result they want to achieve, or they can look also at the impact of their initiative on the people who can make it work. I believe that the latter approach is both instrumentally more effective and the only ethical approach that looks at researchers not merely as means—human resources from which to extract value—but also as ends in themselves, as people whose life can be made better or worse. I therefore urge promoters of DKRs to use the concept of expediential trust in order to adopt the point of view of researchers and take into account their actual working conditions.

If promoters want to raise the chances that scientists share their data, they should better ask themselves what kind of support, compensations and rewards will make it possible for them to share, while keeping their workload manageable, and their life-work balance healthy and sustainable, while achieving their professional goals and not losing competitive advantage. In the past decades we have seen digital technologies bringing many advantages to users while making the conditions for the providers of goods and services worse. Surely, we do not want science to do that to scientists, especially if we believe in the strategic importance of knowledge and that “a motivated academic workforce, satisfied with their reconstructed academic jobs, is most likely to produce the greatest benefit to research, innovation and society” (Bentley et al. 2013, 240). Perhaps it is time to turn on its head the famous exhortation by president Kennedy⁶ and not to ask what scientists can do for science, but what science policy and scientific projects can do for scientists as workers already experiencing heightened demands and declining well-being. If we want to do science as if researchers mattered, then to think in terms of expediential trust when we design digital repositories would help.