Keywords

1 Introduction

In the last decades, societies around the world have been confronted with global phenomena such as inequality, climate change or pandemics and their negative consequences for people and the planet. These and other related issues are commonly known as Grand Societal Challenges (GSCs), which are complex, uncertain, and evaluative (or value-laden) problems (Ferraro et al., 2015) which, to be addressed, require that public, private, and non-profit actors join efforts and pool resources in collective action (George et al., 2016; Gümüsay et al., 2022; Voegtlin et al., 2022). Specifically, addressing GSCs often calls for the collaborative development of innovative solutions, as maintaining or improving the population’s living standards challenges societies to “create, implement and diffuse new products, processes, and services” that do not (further) damage the Earth’s life-support system (Voegtlin & Scherer, 2017, p. 227). Innovation becomes particularly critical when GSCs emerge abruptly and unexpectedly: the COVID-19 pandemic and the challenge of developing, producing, and distributing reliable tests and vaccines are a case in point (Howard-Grenville, 2021; Scherer & Voegtlin, 2020).

In this context, Responsible Innovation (RI) represents a promising avenue for multi-actor, multi-level collaboration aimed at addressing GSCs (Blok, 2019; Owen et al., 2013; Stilgoe et al., 2013; Voegtlin & Scherer, 2017). The concept of RI was originally used in risk-assessments of scientific innovations but evolved to be applied to the general assessment of the ethical and social implications of innovations (Owen et al., 2013). Seen from this perspective, it describes the “transparent, interactive process by which societal actors and innovators become mutually responsive to each other with a view on the (ethical) acceptability, sustainability and societal desirability of the innovation process and its marketable products” (von Schomberg, 2012, p. 50). As an outcome of this process, innovations should be developed that (1) do no harm and (2) do good to people and planet. Further (3), and as a necessary condition for both (1) and (2), their development should also be responsibly governed (Scherer & Voegtlin, 2020; Voegtlin & Scherer, 2017). In particular, acknowledging the tensions that can arise when multiple actors from various domains with different interests and competences collaborate in addressing GSCs, and how these tensions could and should be managed in the innovation process, represents a critical issue for responsible governance.

Recent research on RI and GSCs has highlighted that deliberation, defined as “debate and discussion aimed at producing reasonable, well-informed opinions in which participants are willing to revise preferences in light of discussion, new information, and claims made by fellow participants” (Chambers, 2003, p. 309), can help “balance governance tensions between seemingly opposing governance directives in the effort of integrating stakeholders and allowing for reflexivity” (Voegtlin et al., 2022, p. 11; see also Dryzek & Pickering, 2017). Consequently, fostering deliberation in the governance of innovation can facilitate the development of responsible innovations that legitimately and effectively address GSCs (Scherer & Voegtlin, 2020; Voegtlin et al., 2022). To this purpose, deliberation itself should be fostered through the establishment of deliberative capacities.

Building on Chambers (2003) and Dryzek (2009), Voegtlin et al. (2022, p. 12) define deliberative capacities as the “structures and processes” that support deliberation by allowing to “draw[] on the exchange with those having a stake in the innovation process and outcome with the aim of producing reasonable and well-informed opinions.” Deliberative capacities that are embedded in the formal and informal structures of the innovation process, for instance, provide various arenas for communication and allow for stakeholder participation in networks of complex innovation governance (Fung, 2006, 2015). By fostering deliberation, deliberative capacities help balance the tensions that can arise from the inclusion of multiple actors and perspectives in innovation governance: tensions between the needs (1) to include both expert knowledge and public participation as sources of knowledge in the innovation process, (2) to moderate public discourse between opening up for diversity and closing down for consensus on innovative solutions, (3) to find a balance in the institutional design between centralized, poly-centralized, and de-centralized decision-making, and (4) to address both flexibility and stability in the institutional dynamics and their innovative outcomes (Dryzek & Pickering, 2017).

Yet, while existing research makes the case for the usefulness of establishing deliberative capacities to govern innovation in a responsible way (Scherer & Voegtlin, 2020; Voegtlin et al., 2022), this represents a necessary, but not yet sufficient condition to develop innovations that meet the three criteria of RI listed above. The reason for this is that deliberation is not inherently sufficient for making better collective decisions. While the reasonable nature of a debate and its participants, as envisioned by Chambers (2003), is a necessary component of good deliberation, a debate along these lines could still exclude relevant stakeholders. Acknowledging this, we believe that the concept of communicative reason developed by Habermas (1990, 1995, 2022), which posits that both inclusion and argumentation are normative conditions for the discursive justification of social and moral norms, more closely reflects what can be considered good deliberation. Even more so, inclusion and argumentation are still not sufficient for making good collective decisions until they have real impact on and consequences for public policy. Building on Habermas again, we argue that inclusion and argumentation have “use-value” for stakeholders only insofar as they result in concrete and binding collective decisions that benefit the people in protecting their social, ecological, and cultural rights (see Habermas, 2001, p. 77, on the role of citizens in democratic governance). Therefore, as we argue in this chapter, we need to unpack the concept of responsible governance and suggest that deliberative capacities should meet a number of qualifications to foster deliberation that can be deemed good and helpful in governing innovation that is responsible. Yet, bringing the quality of deliberative capacities to the foreground also implies acknowledging that several challenges may complicate these capacities’ adequate functioning and hence alter their usefulness in supporting responsible innovation governance. In this chapter, we therefore address the questions of how should the quality of deliberative capacities be assessed, what are the challenges to responsible governance in the context of GSC, and how can and should deliberative capacities be improved?

To address these questions, we draw on the literature on deliberative democracy (e.g., Dryzek, 2009; Dryzek et al., 2019; Habermas, 1998) which originally developed in political science and is to date the most influential approach in conceptual and empirical work on collective decision-making (Dryzek et al., 2019; Fishkin, 2018; Fung, 2006) and has also been applied to management studies (Moon et al., 2005; Palazzo & Scherer, 2006; Sabadoz & Singer, 2017; Scherer & Palazzo, 2007). In the literature, a system’s deliberative capacity is defined as “the extent to which a […] system possesses structures to host deliberation that is authentic, inclusive and consequential” (Dryzek, 2009, p. 1382). This means that deliberative capacities, understood as structures (and processes) for collective decision-making, should support a system’s deliberative capacity by ensuring that deliberation (1) “induce[s] reflection noncoercively, connect[s] claims to more general principles, and exhibit[s] reciprocity” (ibd.), (2) allows for including stakeholders (i.e., all who can affect or are affected by the outcomes of decision-making) in deliberations such that their voices are heard and their interests, knowledge, and arguments are taken account of, and (3) consequently leads to public policies, actions, and outcomes that (for the time being) avoid harm and do good to the respective collective, or in the case of GSC, to people and planet more broadly.

Using the development of the SwissCovid exposure notification application during the COVID-19 pandemic as an illustrative case, we propose how each of these three criteria (authenticity, inclusiveness, and consequentiality) could be used to assess the quality of deliberative capacities established for the governance of responsible innovation. Further drawing on this illustrative case, we show how challenges to deliberation can arise in the governance of innovation processes. Finally, we suggest ideas as to what could be done to mitigate these challenges. In line with Frémeaux and Voegtlin (2023), we propose that in the context of democratic nation states, concerned actors should have the freedom to deliberate and device solutions, yet higher-level authorities should be ready to intervene if the quality of deliberation is endangered, for instance by the behavior of some actor(s). However, this presupposes that strong regulatory and enforcement mechanisms are in place at the institutional level so that central authorities are willing and able to intervene legitimately and effectively. As our example shows, this was only partially the case in the development of the SwissCovid and other national exposure notification apps. Further, actors should make sure that the information available to support the deliberation is not only transparent, but adapted to the level of complexity of the issue at stake and the level of knowledge of affected stakeholders. Finally, actors should seek to establish trust between deliberation participants as a way to support open discussions, yet ensure that this trust arises mutually and voluntarily, and not because no other viable option is available (e.g., triggered by dependency on a powerful actor).

This chapter therefore makes three contributions to the literature on RI governance in the context of GSCs. First, it develops criteria to assess the quality of deliberative capacities in innovation governance. Deliberative capacities are embedded in the organizational governance and the governance structures of the surrounding democratic institutions on local, national, regional, or global levels. Depending on their quality they facilitate (or impede) deliberations among multiple actors that are inclusive, authentic, and consequential. They help manage the various tensions in innovation governance and allow for reflexivity in collective decision-making, e.g. taking appropriate corrective measures depending on the performance (Dryzek & Pickering, 2017). Second, our chapter outlines boundary conditions which deliberative capacities should meet to be conducive to deliberation that is inclusive, authentic, and consequential. As literature on RI governance has so far focused on showing how deliberative capacities potentially contribute to solving governance tensions, we contribute with our chapter to the need to better understand the conditions under which such capacities effectively do so. Third, we illustrate our theoretical contribution with the help of a case study on the development of the SwissCovid App, thereby demonstrating not only the theoretical viability, but also practical usefulness of our assessment framework (see Chap. 11 for an Italian case of collaborative response to GSCs).

2 GSCs, Responsible Innovation, and Deliberative Capacities

GSCs are characterized by their complex, uncertain, and value-laden nature (Ferraro et al., 2015). They are complex, in that they affect multiple actors across multiple locations; yet, individual actors also comprehend them in multiple ways, as they are unable to develop a complete and true overview of their causes and consequences (Schneider et al., 2017; Sterman, 2001). This complexity is enhanced by GSCs’ dynamic nature, and the interplay of multiple factors (Ferraro et al., 2015). Second, GSCs are uncertain, as they are mostly non-linear and their future states and linked probabilities cannot be assigned (Voegtlin et al., 2022). Finally, GSCs are also value-laden: individual actors not only are unable to develop a complete picture of the factors underlying GSCs and their interplay, but they also interpret and evaluate this information differently depending on their interests, worldviews, and value systems (Ferraro et al., 2015; Gümüsay et al., 2020).

These characteristics of GSCs impact how innovation aimed at addressing them should be governed. Specifically, as Voegtlin et al. (2022) argue, it is essential that the structures and processes underlying the development, production, and distribution of innovations are reflexive, that is, are able to “reflect on [their] performance” and “reconfigure [themselves] in response to such reflection” (Voegtlin et al., 2022, p. 11; see also Dryzek & Pickering, 2017).

As deliberation is central to achieving reflexivity (Dryzek & Pickering, 2017), building deliberative capacities has been proposed as an avenue to deal with the governance tensions inherent to innovation processes. These tensions concern, for instance, the balance between public participation and expertise, opening up for diversity or closing down for consensus, centralization, and decentralization in collective decision-making or flexibility and stability (Dryzek & Pickering, 2017; Voegtlin et al., 2022), issues that typically arise when innovation necessitates the involvement and collaboration of actors from multiple backgrounds with fragmented knowledge bases and diverse interests and world views (as is often the case to address GSCs). In this context, establishing deliberative capacities at the organizational (Scherer & Voegtlin, 2020) or institutional (Scherer & Palazzo, 2011) levels can “create slack resources that can selectively either open up or close down the innovation process for stakeholders” (Voegtlin et al., 2022, p. 12).

Concretely, research proposes among others three specific avenues for building such deliberative capacities for RI: (1) establishing stakeholder networks, (2) providing arenas for regular exchange and rules-based discourse, and (3) accumulating trust (Blok, 2019; Voegtlin et al., 2022, p. 15). Stakeholder networks can and should be leveraged to provide either broad inputs or specific knowledge into the innovation process, therefore granting flexibility and agility to address the complex, uncertain, and value-laden nature of GSCs. Similarly, establishing arenas for regular exchange and rules-based discourse allows the initiators of an innovation process (for instance, managers in an organization or administrators in public authorities) to selectively draw on a variety of communication and deliberation channels along the innovation process. Stakeholder networks and arenas for exchange and discourse both allow initiators of innovation processes to engage in deliberation with stakeholders about, among others, the goals and legitimacy of the innovation process and its outcomes, as well as the means to effectively and efficiently reach these goals (Scherer & Voegtlin, 2020). The accumulation of trust can in turn be seen as a consequence of well-functioning stakeholder networks and discourse arenas, as both help build the trust that is necessary for the various actors to work together despite their diversity of values and worldviews (Gümüsay et al., 2020). Figure 14.1 below provides a simplified framework of how these factors theoretically interact to support deliberation.

Fig. 14.1
A text box of deliberative capacities has stakeholder arena and arenas for exchange together leading to trust, which points to good interrogation mark deliberation, which is outside the box.

Framework of relationship between deliberative capacities and deliberation as discussed in the RI literature

Full size image

Yet, the usefulness of such networks and arenas depends on the extent to which they are conducive to good deliberation, an issue that has so far not been considered sufficiently by studies on the governance of responsible innovation. Scholarship on deliberative democracy (Dryzek, 2009; Dryzek et al., 2019; Habermas, 1998) has addressed this question by stating that a (political) system should possess capacities which support deliberation that is “authentic, inclusive, and consequential” (Dryzek, 2009, p. 1382). Authenticity implies that “deliberation must induce reflection non-coercively, connect claims to more general principles and exhibit reciprocity” (Dryzek, 2009, p. 1382; see also Dryzek, 2000). Inclusiveness in turn refers to “the range of interests and discourses present in a political setting” (Dryzek, 2009, p. 1382). In particular, being inclusive means allowing for “the inclusion of diverse stakeholder voices and the possibility for public participation” (Voegtlin et al., 2022, p. 16). Finally, consequentiality means that “the deliberative processes must have an impact on collective decisions and social outcomes” (Dryzek, 2009, p. 1382).

In this chapter, we contend that deliberative capacities for innovation governance should likewise be built, assessed, and developed further with these three criteria in mind. For instance, inclusiveness could be fostered, among others, through the design of stakeholder networks. This means that, when establishing stakeholder networks, particular care should be granted to the diversity and balance of voices, interests and knowledge bases that are represented therein. Literature on RI governance sees the main benefit of stakeholder networks in the flexibility that they provide to draw on various forms of knowledge when necessary (Dryzek, 2009; Voegtlin et al., 2022) instead of engaging stakeholders at any cost. Yet, ideally, the innovation process should be as open as possible so that any potentially affected stakeholder can have a chance to voice ideas or concerns at all relevant stages throughout the process. No stakeholder or group of stakeholders should be excluded from a network on the grounds of different opinions or values, and actors should be open toward each other’s values and worldviews. This also means that stakeholders should be given access to communication channels as well as transparent and understandable information, which we discuss next.

Likewise, arenas for regular exchange could contribute to fostering authentic communication and deliberation. Typically, arenas should be built with a set of rules that govern discourse, making sure that claims are substantiated and formulated in terms others can understand (Dryzek, 2009) and that validity claims can be checked so that a collective decision based on reasons can be made (Habermas, 1995). Further, information that is relevant to decision-making within the innovation process should be made available in a transparent and understandable way, so that participants can form their opinions in a manner that is neither coerced by incomplete, non-understandable, nor false information. Arenas should be designed and based on rules that allow for communication between stakeholders and facilitate the formation of collective decisions in a rational way. Typical examples would be online forums or communities in which contributions are moderated either by a central and legitimate authority or debated de-centrally by other users. Here, the discourse rules as formulated by Habermas (1995) can provide a guideline for the structuring of deliberation; these rules concern “the freedom of access, equal rights to participate, truthfulness on the part of participants, absence of coercion in taking positions and so forth” (p. 56).

Finally, the levels of trust developed between deliberation participants could contribute to making the innovation process consequential, that is, having an “impact on collective decisions and social outcomes” (Dryzek, 2009, p. 1382). For instance, building trust across different stakeholder levels fosters open discussions in the innovation process, increasing the chances that the outcomes thereof are understood, considered legitimate and broadly adopted, also by the broader population which was not directly involved in the process but represented by others that share their views (Fung, 2006, 2015). This is especially relevant for innovations in contested or ethically sensitive areas, such as genetic engineering or artificial intelligence (Voegtlin et al., 2022). The societal adoption of an innovation therefore constitutes a relevant indicator of the consequentiality fostered by established deliberative capacities. Considering the criteria presented above and how they apply to the design and assessment of deliberative capacities, Fig. 14.2 provides an updated normative framework of the relationship between deliberative capacities and deliberation that we propose in this chapter.

Fig. 14.2
A text box of deliberative capacities has stakeholder arena and arenas for exchange together leading to trust. The components of the flow outside the box include inclusiveness authenticity, good exclamation mark deliberation, and consequentiality.

Updated framework of the relationship between deliberative capacities and deliberation

Full size image

As developed above, building deliberative capacities is a necessary, but not sufficient, condition for deliberation to take place that is inclusive, authentic, and consequential. Rather, such deliberative capacities should be designed in a specific way so that deliberation meets these criteria. In some cases, however, factors that are inherent to stakeholders involved in the innovation process (e.g., unwillingness to change positions) or to the issue toward which innovation is targeted (e.g., issue is complex and demands immediate action thus leaving too little or no time for deliberations) may challenge deliberation.

In the next section, we build on the case of the development of COVID-19 exposure notification apps to provide illustrations of such challenges and how they affect a system’s deliberative capacity.

3 Challenges to Deliberation: The Case of the SwissCovid Exposure Notification App

The COVID-19 pandemic is generally considered in the literature as a GSC (see, e.g., Howard-Grenville, 2021; Schwoon et al., 2022; Voegtlin et al., 2022). Its complex, uncertain, and value-laden character is exemplarily illustrated by, for instance, the trade-offs between related issues such as the safeguarding of public health and the preservation of the economy, or the unpredictable mutations of the virus and the emergence of social movements trivializing the virus and criticizing or resisting government measures adopted to halt its spread.

Along with vaccines, exposure notification apps represent one of the most significant and ethically challenging innovations that emerged during this period (Morley et al., 2020; Parker et al., 2020). Initially conceived as a responsible innovation that would limit the spread of the virus and allow the mapping of transmission by health authorities, the debate around these apps quickly crystalized around the potential harms that such an innovation could pose to the privacy of individuals (Sharon, 2021; Zuboff, 2022). The core issue was that notifying individuals of their exposure to (potentially) infected people would require tracing (although in a theoretically anonymous way) whom they had been in contact with over a more or less long period (generally 2 weeks). Building on the case of the development of the SwissCovid exposure notification app in Switzerland, we illustrate the challenges that may arise at various levels in innovation processes despite the presence of (moderate) deliberative capacities.

3.1 Case Background

In Europe, the development of an exposure notification protocol was initiated in March 2020 by a multinational coalition of European scientists and technologists as the Pan-European Privacy-Preserving Proximity Tracing (PEPP-PT) (Abboud et al., 2020). Initially, the project aimed at supporting both centralized and de-centralizedFootnote 1 approaches to Bluetooth-based exposure notification. In this context, “centralized” meant that IDs generated by users’ smartphones are “uploaded to a trusted server, such as one controlled by a health authority,” while “de-centralized” meant that “IDs are held locally on devices, where the infection risk is also calculated” (Lomas, 2020b, para. 11). However, due to concerns raised by members of the original group of scientists about the transparency of the PEPP-PT approach and conflicting views on the issue of privacy and the risk posed by state authorities centrally managing data, teams quickly ended up splitting, and a dissenting fraction started working on another project which would follow what they called as a “de-centralized” approach only, the De-centralized Privacy-Preserving Proximity Tracing (DP-3T) (Powers, 2020b).

The atmosphere between both teams became tense (Abboud et al., 2020; Powers, 2020b), with the DP-3T team, for instance, accusing the PEPP-PT team of trying to deceive the scientific community and the broader public about the harmlessness of their approach and its implications of privacy concerns. As one scientist involved in the DP-3T stated:

We published the DP-3T protocol in early April for discussion and feedback, but it soon became apparent that PEPP-PT was building a Trojan horse: using the privacy community’s wide approval of our public system to slip their own, unpublished centralized approach into deployment. (Veale, 2020b, p. 36)

Around the same time, on April 10, technology giants Apple and Google made the surprising announcement that they were collaborating to develop an Application Programming Interface (API) based on similar principles as the DP-3T protocol. This API would allow exposure notification apps developed by national health authorities to use Bluetooth in the background, an essential prerequisite for the app to function without draining too much smartphone battery (Veale, 2020b). It would, however, embed contact tracing technology in the operating system layer of smartphones running on Android and iOS, instead of at the app (or software) level. This raised concerns about the creation of a “dormant functionality for mass surveillance at the operating system layer” (Hoepman, 2021, p. 10) which would be under the control of the two technology firms rather than a public health authority. Despite these concerns, the European Parliament and the European Data Protection Board, as well as several European countries, including Switzerland and the United States, quickly expressed their preference for the “de-centralized” solution powered by Apple and Google (Sharon, 2021; Veale, 2020b; Zuboff, 2022). Apps based on the centralized solution, less efficient because they were not granted access to background Bluetooth by the Apple-Google API, were only pursued by a few countries (among which France, the UK, and Australia), while the original PEPP-PT consortium eventually collapsed (Veale, 2020b). The SwissCovid app was officially launched on June 25th, 2020. While it was downloaded over a million times, it, and other national apps, did not have the expected uptake, due mainly to a lack of trust from the public (Blasimme & Vayena, 2020; Zuboff, 2022).

In the following section, we assess the quality of the deliberative capacities involved in the development of the SwissCovid app and discuss challenges that possibly diminished the quality of deliberation during the innovation process. We argue that the main challenges to good deliberation arose from actors’ behavior, issue complexity and lack of mutual trust, leading in turn to low public willingness to endorse the innovation outcome.

3.2 Challenges to Deliberation

Challenges to Inclusiveness–Actor Behavior

As highlighted above, stakeholder networks should provide the grounds for deliberation that is inclusive. At the outset, the SwissCovid innovation project was conceived as a collaborative network of scientists, firms, and national authorities. While the public was not directly involved in the project, but only represented via their (democratically elected) government officials, documents such as the code underlying the app were stored on GitHub, an open internet platform which allowed to interact with the project developers (more on this in the next section). A first challenge to inclusiveness emerged when scientists from the PEPP-PT and DP-3T teams started arguing about whether a centralized or a de-centralized solution would be better to safeguard the privacy of the users. As the two endeavors finally split following disputes around privacy and the PEPP-PT removing any mention of the de-centralized protocol from its website without telling the DP-3T team beforehand (Powers, 2020b), the DP-3T team initiated a communication and lobbying campaign against the PEPP-PT team working on what they called the “centralized” solution (Vaudenay, 2020). As an observer notes, the objective of this campaign was to influence policymakers to favor one option over the other, excluding the PEPP-PT team from the broader endeavor without engaging in any real academic discussion about centralized versus de-centralized systems (Vaudenay, 2020, p. 2).

The second and most salient challenge to the inclusiveness of deliberation was triggered by the behavior of the two most powerful actors in the network, namely Apple and Google. Thanks to their dominance over the market for smartphone operating systems, both actors also possessed an enormous amount of power over the innovation process as any resulting app would need to run on both Android and iOS. In April 2020 they announced “with little, if any, communication with local experts and EU policymakers” (Scott et al., 2020, sec. 2 para. 3) that they would launch their own solution, which not only would embed contact tracing directly in the operating system, but would also support de-centralized protocols exclusively (Hoepman, 2021). Consequently, both firms unilaterally refused to adapt their operating system to also make it work with centralized protocols such as the ones developed with support by the democratically elected governments in France and in the UK (Veale, 2020a).Footnote 2 Instead, the two technology firms publicly advocated for apps running on a de-centralized model, such as the SwissCovid app. This move allowed both Apple and Google to set the terms and conditions of what contact tracing should be and how it should work (Hoepman, 2021), yet, without sufficient democratic entitlement and control (Zuboff, 2022). By design, both actors excluded democratic countries whose health authorities favored a centralized solution to access the nature of contacts for epidemiological reasons (i.e., to build a model of contacts to map the spread of the virus) (Hoepman, 2021). The Australian Minister for Government Services, for instance, stated that the Australian Government, who pursued a centralized solution, was open “to improving [the] technology” if such technology maintained a key role for health officials in the process, adding that “the current structure of the Google-Apple API does not do that” (Bogle, 2020). Both firms even refused to engage in reflection with those teams (Zuboff, 2022). As a consequence, the teams of several countries exploring a centralized solution had to abandon their efforts and eventually adopt the de-centralized protocol (e.g., the UK), and those who continued (e.g., France, Australia) did so by sacrificing the operability, and hence the attractiveness and efficacy of their solution.

Challenges to Authenticity: Actor Behavior and Issue Complexity

The power and behavior of the actors involved in the innovation process also presented a challenge to the authenticity of deliberation, which presupposes that claims are substantiated and formulated in terms that others can understand and accept, and induce reflection noncoercively, so that collective decisions can be made based on reasons and shared insights (Dryzek, 2009). In the present case, although the innovation process was meant to be open source and the code underlying the DP-3T protocol and the SwissCovid app was developed in a co-construction process, Apple and Google refused to publish the source for their API code until it was fully developed. Following this move, actors involved in the innovation process, and those simply interested in it, were able neither to scrutinize the development of the API nor to run the codes of other apps based on the DP-3T protocol. This negatively impacted actors’ ability to make informed decisions. The conditions both firms laid out for supporting national apps (requiring them to be built around a de-centralized design) further acted as a form of coercion, silencing potential discussions around alternative designs and other non-privacy-related issues (more on this in the next paragraphs). As critics noted, the attitude adopted by both firms led to “citizens hav[ing] no longer anything to say about it (the app), except taking it as a whole or refusing it” (Vaudenay & Vuagnoux, 2022, sec. 3, para. 5–6).

The negative impact of actors’ behavior on the authenticity of deliberation was enhanced by the characteristics of the issue at stake, in particular its complexity and urgency. As introduced above, in the case of the SwissCovid app development, the main arena for exchange during the innovation process was the GitHub platform on which the source code was stored (Vaudenay, 2020). The platform is commonly used to host open source software development projects, as it allows registered users to follow the development of a software and interact with its developers (Finley, 2015). Yet, due to the complexity of the issue, the existence of this exchange arena was not sufficient to trigger authentic deliberation. As understanding the specificities of the centralized and de-centralized protocol designs required an understanding of coding language and technicalities, the majority of the lay population was unable to form their opinion in a fully informed way (although some scholars attempted to explain the risks inherent to all exposure notification solutions in a simplified way; see, e.g., Bonnetain et al., 2020). Most interested citizens therefore had no choice but to revert to the broader public sphere (the media, political debates, etc.) to find information on the application and its development.

However, the discourse in the public arena was mostly incomplete and imprecise. Historically, combatting epidemics had strongly relied on the work of “public health ‘surveillance systems’,” which were supported by “individual case data in some combination with epidemiological statistical tracking” (Zuboff, 2022, p. 44). Yet, Apple and Google’s involvement in the development of digital contact tracing quickly brought the need to protect individual privacy at the center of discussions, at the expense of “the public health imperative of tackling a virus” and other issues (Scott et al., 2020, sec. 1 para 4; see also Sharon, 2021; White & van Basshuysen, 2021). Surprisingly, political authorities also promptly adopted this privacy focus: as early as April 17, 2020, the European Parliament demanded in a resolution that apps follow a de-centralized protocol (Veale, 2020b) despite the proven effectiveness of centralized data storage for contact tracing (White & van Basshuysen, 2021). This incomplete debate was also plagued with imprecise information: Serge Vaudenay, a professor from the Security and Cryptography Laboratory of the Swiss Federal Institute of Technology (EPFL), for instance, criticized that available information concealed critical issues such as the privacy concerns also emerging from the so-called de-centralized solution (e.g., Vaudenay, 2020; Vaudenay & Vuagnoux, 2022). Overall, the arguments against centralized solutions used in the information to the general public were criticized as “overly exaggerated and the ones in favor of de-centralized systems have been oversold” (Vaudenay & Vuagnoux, 2022, sec. 1, para. 3).

Finally, the public discourse was also criticized as being emotionally laden, as the rhetoric of “centralized” vs. “de-centralized” was used in a way that instilled fear of an untrustworthy state that would surveil citizens, while “the image of ‘decentralization’ immediately conjured associations with pro-social democratic ideals” (Zuboff, 2022, p. 46). As an illustrative example, a professor working on the DP-3T proposal once described a centralized solution as potentially “open[ing] the gates to privacy hell” and “becom[ing] a wet dream for security services” (Powers, 2020a, para. 12) as it could allow governments to find out who is in close proximity to whom. Altogether, the incomplete and imprecise nature of the public debate around contact tracing further coerced most stakeholders in the innovation process into adopting the de-centralized solution.

The urgency of the issue also led various stakeholders to abandon the ideal of deliberation and instead align with the requirements of Apple and Google without any further questioning or scrutiny. Pressured by the need to develop and rollout a solution that would reach the levels of adoption and interoperability necessary for it to be effective, many national governments acquiesced to the requirements of those who held the keys to this new public good: Apple and Google (Sharon, 2021; Zuboff, 2022). Scott (2020, sec. 2 para. 8), for instance, reported the words of a German government official, who stated that “we need to have a discussion on how Silicon Valley is increasingly taking over the job of a nation state […]. But we don’t need to have it amid a pandemic.”

As this case shows, the complexity and urgency of the issue and the resulting incapacity of most citizens to understand its stakes paved the way for views on individual privacy and government trustworthiness to be expressed using highly emotional rhetoric, something the literature on deliberation regards critically (Dryzek, 2009). In other words, this example illustrates how issue complexity and urgency can act as further challenges to deliberation that is authentic. An overly complex issue and the lack of understandable information formulated in a language that everyone can understand impedes de facto that actors, in particular, the lay public, make an informed and noncoerced decision as to whether or not an innovation meets criteria of avoiding harm and doing good. These challenges had implications for the building of trust between stakeholders and deliberation consequentiality, as highlighted in the following section.

Challenges to Consequentiality: Actor Behavior, Issue Complexity, and Mutual Trust

As the RI literature states, trust is an essential component of deliberative capacities (e.g., Voegtlin et al., 2022). Trust is commonly defined in the organizational literature as “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” (Mayer et al., 1995, p. 712). However, for good deliberation to occur, trust should be mutual, so that actors are willing to open up to each other and strive for a “meta-consensus,” or “agreement on the legitimacy of disputed values, the credibility of disputed beliefs, the nature of disputed choices (including the range of acceptable options), and the acceptable range of contested discourses” (Dryzek & Pickering, 2017, p. 357; see also Voegtlin et al., 2022).

In the SwissCovid app case, issue complexity and the strategic behavior of actors complicated establishing trust that is mutual and conducive to good deliberation. While all parties agreed on the importance of preserving user privacy, the range of acceptable options, in this case the choice between a centralized or a de-centralized design, was complex and subject to heated and emotion-laden discussions. Promoters of a de-centralized design (including Apple and Google) were not willing to rely on national governments to store and manage contact data, while promoters of the centralized design were reluctant to rely on contact tracing by commercial actors (e.g., Sharon, 2021; Veale, 2020b; Zuboff, 2022). The dispute was enhanced by the fundamentalist character of some stakeholders not “interested in the kind of reflection upon values and beliefs that is central to deliberation” (Dryzek, 2009, p. 1396). The libertarian and anti-state ideology that dominated in the Silicon Valley and among many computer scientists was typically at the center of the clashes (Zuboff, 2022) that resulted in the failure to reach a meta-consensus (Dryzek & Pickering, 2017) and proponents of the de-centralized solution trying to “shut down” the effort of their former colleagues (Abboud et al., 2020).

As the app was finally developed in collaboration with Apple and Google, it is worth noting that developers’ willingness to rely on the solution supported by both firms was driven by its technically inevitable character, as the companies control the operating systems running on almost all smartphones. Specifically, as the scientific community found itself trapped between the urgency of developing a solution to tackle the spread of the pandemic, and fear of data misuse by malicious governments, many scientists perceived relying on Apple and Google as a necessary evil (Veale, 2020a; Zuboff, 2022). As several experts (including those involved in the DP-3T project), for instance, noted, because Apple and Google implemented the exposure notification at the operating system layer and hence controlled which government apps have access to it, there was a perceived and seemingly inevitable need to trust that both companies “will not abuse [the API]” (Hoepman, 2021, p. 12; see also Powers, 2020c; Troncoso et al., 2020). In other words, while developers showed a certain level of trust by accepting to be vulnerable to Apple and Google, this trust was driven by external constraints and certainly did not suffice for all involved stakeholders to engage in open discussions.

3.3 How Challenges Impact Deliberation Quality

As a consequence of the challenges posed by issue complexity and actor behavior, and despite the presence of some deliberative capacities, the stakeholders involved in the innovation process did not manage to ensure inclusive and authentic deliberation in the governance of the innovation process. This certainly also affected its consequentiality. Although it is complicated to draw a clear causal link between the innovation process and the final adoption of its outcome, what was observed is that the general public also showed little confidence in the application and its ability to protect individual privacy, resulting in relatively low adoption levels in Switzerland and other European countries which implemented similar products (Blasimme & Vayena, 2020; Zuboff, 2022).

To summarize and getting back to the question asked in the title of this chapter, “Is SwissCovid a Responsible Innovation for a Grand Societal Challenge?”, the discussion presented in this paper leads us to answer negatively. In the present case, the deliberative capacities in place did not allow to balance the different tensions that arise in the governance of RI innovation (Dryzek & Pickering, 2017). As a result, conditions were not met for inclusive, authentic, and consequential deliberation to take place. In the following paragraphs, we show how each tension failed to be properly managed, and in the next section we will present potential solutions to improve the deliberative capacity of innovation governance.

A first tension, as identified by Dryzek and Pickering (2017), arises from the need to include both expert knowledge and public participation as sources of knowledge in the innovation process. In the present case, there has been no public consultation, as all decision-making pertaining to contact tracing took place among scientists, politicians, and corporate actors. Yet, interestingly, a public deliberation experiment conducted in Australia showed that “informed citizens are willing to trade their privacy for common goods such as COVID-19 suppression” (Degeling et al., 2022, p. 97), suggesting that the outcomes of the innovation process could have been different would the public have been involved.

The second tension relates to the need to moderate public discourse between opening up for diversity and closing down for consensus on innovative solutions. This tension too failed to be properly managed. While the innovation process was initially open and considered both centralized and de-centralized protocol designs as well as concerns of epidemiological effectiveness next to privacy issues, very quickly this turned to a non-debate as a result of actors’ behavior and the urgency of the issue. Following Apple and Google’s demands and the constraints they imposed, in a matter of days almost all national governments rallied behind the de-centralized solution, thereby closing the debate on alternative solutions and, as some critics pointed out, ignoring critical voices (Vaudenay, 2020; Vaudenay & Vuagnoux, 2022).

The governance of the SwissCovid and other contact tracing apps development also fell short of balancing between centralized, poly-centralized, and de-centralized decision-making. While the European Commission initially granted national governments the freedom to choose between centralized and de-centralized designs (Lomas, 2020a), only requesting that apps be able to communicate, on the next day the European Parliament publicly endorsed the de-centralized design (Veale, 2020b). By setting the conditions for contact tracing apps to work on their operating systems (Tretter, 2023), Apple and Google definitively, and de facto unilaterally, tilted the balance toward centralized decision-making, yet not at the level of democratic political authority, but at the corporate level.

Finally, balancing the tension between the need to address both flexibility and stability in the innovation process and its innovative outcomes would have required that app developers reflect critically on the app’s performance and if necessary debate on the need to adapt it (flexibility) or continue with the existing design (stability). In Switzerland, however, despite relatively low adoption and frequent criticism on its usefulness and safety, no formal efficiency evaluation criteria were developed (Vaudenay & Vuagnoux, 2022). Interestingly, as the app was taken out of service 2 years after its introduction, the critical self-reflection of involved actors did not attribute its mixed performance to design or efficiency flaws, but to the lack of app promotion by national health authorities and the lack of digitalization of the Swiss health system (Buchmann & Tschirren, 2022).

To summarize, the SwissCovid case shows the role played by large power differentials between actors in a stakeholder network, and how the complexity of the issue at stake and timing constraints may enhance the detrimental effects of power differentials. Actors with disproportionate power, such as big corporations, might hijack deliberative capacities to drive the innovation process in a direction that favors their own interests and avoid having to engage in open deliberation and face potential criticism. The institutional conditions that have paved the way for Apple and Google to behave as they did in this innovation process, and the societal and political consequences thereof reach well beyond issues of innovation governance. Researchers (e.g., Veale, 2020b; Tretter, 2023) have, for instance, discussed the implications of both firms’ exclusionary attitude on the sovereignty of nations and their democratically elected governments to administer matters of public health. By refusing to support nations pursuing a centralized solutions, Apple and Google interfered among others with the ability of health authorities to collect epidemiological data, impacting their management of the pandemic (White & van Basshuysen, 2021). Similarly, in a paper on the increasing inference of big tech corporations in public health issues, Sharon (2021) draws on political philosopher Michael Walzer’s theory of spheres of justice (1983) to contend that Apple and Google’s use of their dominance in the digital technology sphere to build dominance in the political sphere equates a form of tyranny. Building on our findings and these reflections, in the following section we propose that deliberative capacities should be extended beyond the focus on organizations to also address the institutional environment in which innovation governance takes place.

4 How to Overcome Challenges to Deliberation?

As the above discussion highlighted, deliberative capacities are a necessary but not sufficient condition for achieving “good” deliberation in RI processes. We have identified several challenges that can affect the quality of deliberation, namely actor behavior, issue complexity and urgency, and lack of mutual trust. In this section, we now outline potential solutions to overcome, or at least mitigate, these challenges.

With regard to the risks related to actor behavior, ensuring that no actors or set of actors possesses a disproportionate share of power over other stakeholders is a well-known challenge of authentic deliberation and is difficult to overcome (Curato et al., 2017; Sabadoz & Singer, 2017). With an eye on deliberative capacities at the organizational and inter-organizational level, one approach is to carefully select stakeholders and establish networks with the balance of actors in mind. In that regard, instead of trying to avoid power imbalances, which is illusory, differences in the quality of power should be considered: one stakeholder possessing more economic power is not necessarily an issue if other stakeholders are more powerful in terms of cultural or social resources that are also relevant in the innovation process. This way a modus vivendi of checks and balances may be reached even in situations with disproportionate power. What should however be avoided, as far as possible, are situations of absolute dependence on one stakeholder, as the SwissCovid app case and the dependence on Apple and Google illustrates. This applies in particular when the products and services at stake in the innovation process have a “public good” character (and their withdrawal would therefore cause “public bads”), as is apparent in the case of the COVID-19 contact tracing app. In situations of disproportionate power and dependency, a robust institutional context becomes essential to guarantee the deliberative capacity of innovation governance.

In a recent paper, Frémeaux and Voegtlin (2023) acknowledge the importance of the institutional context in which deliberation takes place and propose that relying more strongly on the principle of subsidiarity might support a system’s deliberative capacity. Subsidiarity means that “a central authority should have a subsidiary function, performing only those tasks which cannot be performed effectively at a more immediate or local level” (Oxford English Dictionary, 2023). That is, in a society, local authorities should be able to deliberate on and device solutions. Only if local coordination does not work, which is the case with global challenges such as the pandemic, higher-level institutions should intervene, again with a focus on developing effective, efficient, and most importantly, legitimate innovations. In this context, enlightened leaders in higher-level institutions could become the custodians of good deliberation (Patzer et al., 2018). In the SwissCovid app case, the European Commissioner, for instance, urged Apple’s CEO to continue dialogue and support all types of designs, including centralized designs such as the one developed in France (Petrova, 2020). However, Apple’s ultimate refusal to support the French solution (Kar-Gupta & Rose, 2020) raises questions on the adequate level at which this subsidiary function should be located. Further, and more concerning, political institutions not only were left powerless to guarantee the interests of their citizens in the matter (Vaudenay & Vuagnoux, 2022; Zuboff, 2022), but also appeared unwilling to enforce the protection of their sovereignty. As one stakeholder involved in the innovation process wrote:

Oddly, it is notable that there has been little appetite to attempt to rectify this situation with the legal obligations that sovereign states have at their disposal; instead reifying the view of tech giants as state-like themselves, diplomatic interlocutors rather than firms operating under national law. Sovereignty was mourned before any of its traditional tools were even reached for (Veale, 2020b, p. 38).

From this perspective, the SwissCovid case highlights the critical importance of strong regulation and enforcement mechanisms as deliberative capacities that should be established at the institutional level to support authentic, inclusive, and consequential deliberation at the organizational, inter-organizational, and above-organizational level. It however also highlights how crucial it is that national and international authorities become aware of their duty to make use of such mechanisms to protect a system’s deliberative capacity down to the organizational level, and of the far-reaching consequences of their failure to do so.

In a system in which robust supporting institutional safeguards are in place to protect against abusive actor behavior, it becomes possible to develop lower-level deliberative capacities that are targeted to the innovation governance process. Such capacities allow to address innovation-specific challenges such as those arising from issue complexity and urgency. As regards the challenges posed by issue complexity, communication should not only be made transparent, but also understandable to the majority of stakeholders. Further, deliberation should, as much as possible, unfold in terms that others can accept, that is, remain factual and not be too much influenced or tempted by rhetoric and emotional considerations. Research on deliberation has shown that the lay public is able to deliberate even with limited information and produce meaningful arguments and results (Dryzek et al., 2020). A learning from the SwissCovid case could be to alternate, or cascade deliberations with different groups of stakeholders, e.g., deliberations with experts and deliberations with broader, public participation. Deliberations involving the broader public could be centered around the ends (what is the ultimate goal of the SwissCovid app and in how far does that goal serve societal needs/the common good?), while deliberations around the means to achieve the ends (in how far is a de-centralized or a centralized solution technically or otherwise able to most efficiently reach the intended goals?) could be conducted primarily among experts. This process could go back and forth, and the focus on linking the deliberative fora should be on condensing and translating any meaningful agreements (Frémeaux & Voegtlin, 2023). This would ensure that trust between stakeholders is built constructively based on dialogue and openness, and not subject to a “need to rely on” given solutions.

The two propositions we outlined offer relevant learnings for responsible innovation in and with business. Deliberation is a key aspect of ensuring efficient, effective, and legitimate innovation, especially when innovation is targeted at grand societal problems characterized by insecurity and value-laden standpoints of various stakeholders and the public. We used the development of the SwissCovid app as a case to illustrate the various challenges that can arise in innovation endeavors despite the existence of deliberative capacities. Our suggestions can inform business practices and policy making on how to mitigate or overcome these challenges. We suggest subsidiarity and the establishment and enforcement of robust institutional safeguards, as a way to preserve mutual trust and ensure that deliberation is authentic, inclusive, and consequential.

To these same purposes, we further suggest alternating and cascading deliberation around the ends and means of innovation as a way to keep both experts and the broader public involved in their areas of competence yet also with a view over the entire process. However, this can only succeed if institutional conditions are created in the form of “new zones of public governance aligned with the values, principles and aspirations of democratic societies and empowered to hold accountable both market and state to the rule of public law” (Zuboff, 2022, p. 54) to ensure that the objectives of responsible innovation do not get lost along the way. Building on these reflections, we would encourage future research to further theorize on and empirically test the ideas presented in this chapter. It would be interesting, for instance, to reflect further on the conditions under which the involvement of higher-level institutions is most successful in preserving good deliberation. What role do factors such as the broader political context, previous relationships between stakeholders and higher-level institutions, or timing, play? Similarly, future research could theorize and empirically test the optimal separation of deliberation on means and ends and the frequency at which both should alternate in different contexts.

5 Conclusion

GSCs such as the recent COVID-19 pandemic reveal the importance of collaborating across stakeholders and beyond worldviews to develop innovative and responsible solutions. As such challenges require that actors from different backgrounds and with different interests join forces in innovation processes, scholarship has acknowledged that deliberative capacities need to be in place to manage the various tensions that can arise in innovation governance. Yet, the presence of deliberative capacities is not sufficient to foster deliberation that is authentic, inclusive, and consequential.

A more realistic view acknowledges that challenges to deliberation can arise that relate to the behavior of actors within stakeholder networks, the complexity and urgency of the issue which the innovation aims to tackle, and a lack of mutual trust between deliberation participants. These challenges may in turn make the governance of the innovation process less reflexive and jeopardize its ability to generate outcomes that avoid harm and do good. Building on the case of the development of the Swiss Covid exposure notification app, we have provided in this chapter an illustration of such challenges and how they can affect deliberation, as well as outlined potential solutions to mitigate their effects. We acknowledge that the identified challenges are certainly not exhaustive and we encourage future research to more closely explore the conditions under which deliberative capacities can support deliberation that is desirable from an RI perspective.