On with critique! The necessity of critique in addressing the political deficits of responsible innovation

ABSTRACT

This article responds to the calls to address the political deficits of RI frameworks and uptakes by taking recourse to approaches inspired by the methods and emancipatory ambitions of critical theory. By outlining the results of three existing studies, important and varying dimension of R(R)I’s political deficits are identified. Despite each study identifying different aspects of the political deficit, they all share a concern for the power relations implicit in RI frameworks and uptakes. Yet, what is lacking in the critical literature is an appreciation of the role of social norms, social relations, and social structures, which is vital to an analysis and criticism of power. This article encourages RI frameworks to engage with critical practices by addressing the role these factors play in creating, exercising, and perpetuating hegemonic forms of power with the aim of greater social emancipation.

KEYWORDS:

• RI
• critique
• critical theory
• participation
• social emancipation

Introduction

The difficulty of addressing the problem of responsibility is increasingly felt across societies worldwide. The global impacts of recent and ongoing crises such as the Covid-19 pandemic and the Russian invasion of Ukraine, have fostered an acute sense of uncertainty felt on the level of the everyday household. Pressing issues are intensifying, such as food and water insecurity, the need for clean energy, high cost of living, a precarious future alongside artificial intelligence, and the perpetual possibility of nuclear warfare. All of this is part of an era of uncertainty catalyzed by the rapid changes in e.g. climate, technology¹, and social life. These issues, and the uncertainty that comes with them, continue to challenge operative concepts of agency and responsibility. In turn this invites critical challenges to conceptual frameworks such as Responsible Innovation (RI)², which assumes that technological innovation has the potential to address pressing problems as long as the innovation processes are managed in a responsible way. This article will challenge RI frameworks to respond to the criticisms of their political deficits by encouraging the adoption of a stronger political dimension which itself emerges and persists through ongoing participatory and critical practices.

While the concept of critique is complex, I understand it here as a dimension of practice that interrupts everyday experiences and ways of thinking about the world in order to expose inherent presuppositions and claims at work in it. While this practice is often considered to be the exclusive work of the theoretician, French sociologists Luc Boltanski and Laurent Thévenot argue that everyday social actors also engage in critique when they find themselves in ‘critical moments’ (Boltanski and Thévenot 1999, 359). These critical moments occur when actions are prevented from being coordinated or harmonious in some way, constituting a ‘break in the course of action’ (Boltanski and Thévenot 1999, 360). This in turn prompts everyday social actors to respond by means of justification (i.e. justifying their claims). This sociological insight is important as it attributes agency to everyday social actors, rather than assuming that they are passive cogs in a machine. In this light, critique is understood as a part of social reality and hence as a dimension of everyday practice. The specific concept of immanent critique, however, is a more theoretical endeavor. While always grounded in its everyday counterpart, I will understand immanent critique as a theoretical undertaking which critically evaluates a framework on its own terms: What are the (ideal) conceptual underpinnings of the framework? Do the concepts correspond with their practical realizations?

In what follows, RI will be challenged to engage with critique in at least two ways: (1) immanently, referring to its inherent conceptual and normative underpinnings and (2) externally, referring to its (lack of) relation to the everyday experiences of the broader public and the local forms of critique that take place on that level. Without critique, RI remains trapped within the limited consciousness of its own possibility. Or, to put it differently, without critique RI’s operative concept of responsibility is caged by its normative and ideological assumptions and the social structures that keep them in place.

The article will be structured as follows: firstly, I provide a brief overview of the role of public inclusion as well as deliberative and participatory practices in RI frameworks. Then, in order to position my argument in relation to the critical literature, I outline the results of three studies that identify political deficits in RI and its uptake. By setting up this critical landscape, I show the tension between the desire for democratic inclusion in the innovation process and the way institutions suggest implementing RI frameworks. Here we see that politics, participatory practices, and critique are often either seen as negative or treated superficially and are replaced by a concept of (innovation expert) management wherein controlled inclusiveness is an important feature. Finally, I engage with the work of critical theorist of technology Andrew Feenberg whose work is often overlooked in the RI debates that deal with its political and critical deficits. Feenberg’s contributions regarding the democratization of technology are constructive and can be used to critique RRI’s operative key concept of ‘inclusion’ to show that exclusionary (technocratic) practices remain implicit in the way these frameworks conceptualize and integrate public participation in the innovation process. Feenberg’s work can be an important tool in the debates on how to conceptualize political and democratic deficits and how to politicize RI. Specifically, I show how a critical theory of technology can offer a more comprehensive conception of political practices in which public participation and critique are integral to the framework. In doing so, I offer a more comprehensive criticism than existing critiques of the political deficits of R(R)I and thus respond to those who have called for the political deficits in RI frameworks to be addressed (Frahm, Doezema, and Pfotenhauer 2022; Ludwig et al. 2022; Owen and Pansera 2019; van Oudheusden 2014; von Schomberg and Blok 2023).

Reflections on public inclusion and deliberative and participatory practices in R(R)I

While the term Responsible Innovation currently appears in many institutions and sectors globally, R(R)I proponent Rene von Schomberg has argued that RI will only be achieved on the following condition: when public and private stakeholders join forces and become mutually responsive to one another, in order to anticipate the research and innovation outcomes needed to responsibly address what the Lund Declaration has identified as the grand challenges of our time (von Schomberg 2013; 2015). Aligning Responsible Innovation with particular (European) values has been criticized given that ‘in different areas of innovation, and in different cultural contexts, different values may be more or less pertinent, and they may be conflicted’ (Stilgoe, Owen, and Macnaughten 2013, 1577). This has led to a more general framework for Responsible Innovation that includes the dimensions of anticipation, reflexivity, inclusion, and responsiveness without explicitly defining RI’s normative ends (Stilgoe, Owen, and Macnaughten 2013). Nonetheless, the dimension of inclusion is significant for RI’s political self-conception (e.g. democratic, participatory) as ‘the category of inclusiveness, … also implies a high priority for participatory processes’ (Grunwald 2019, 705). Indeed, also for von Schomberg, the dimension of inclusivity is central to RI, specifically in its aim to include various actors throughout the innovation process; inclusion is key to a successful Responsible Innovation framework. However, despite the seeming importance of inclusion, RI has been criticized for its deficient understanding of the related concepts of inclusion, participation, and (deliberative) democracy as well as how that understanding plays out in practice (in various uptakes of RI).

The way these concepts are operationally used in RI frameworks assume that more inclusion and more participation will necessarily lead to more democratic outcomes.. There are traceable historical explanations for why this is the case. For instance, the social and political climate in Europe in the late 1990s led to growing distrust in European institutions, ‘experts’, and politics in general (Owen, von Schomberg, and Macnaghten 2021, 218). In the early 2000s ‘science and technology development were seen as being central to the knowledge economy and European competitiveness, and in this regard securing societal acceptance of new technologies and trust in science were deemed essential’ (Owen, von Schomberg, and Macnaghten 2021, 218, emphasis mine). This key incentive led in part to the development of the RRI framework by which the European Commission tried to shorten the distance between society and science and technological innovation. This in turn led to the explicit thematization of the role of public engagement and deliberation in research and innovation. Yet the way which these concepts are operationalized remains narrow. This is because RI frameworks disregard the need for critical processes not only to detect biases (e.g. within those deliberative and participatory contexts) but also to tap into the critique of everyday social actors that struggles to find articulation in uptakes of RI frameworks.

To provide a concrete example, RI frameworks have been criticized for their lack of critical reflection on (1) how deliberative processes are carried out in practice and (2) the conceptual – and biased – building blocks that constitute a given deliberative context (van Oudheusden 2014). Further, given that ‘Global and Grand Challenges’ [GGCs] – prominently used in RI discourses – require innovative solutions, critical calls have been made to address the lack of attention ‘to the question of how responses to GGCs are negotiated’ (Ludwig et al. 2022, 7). Studies show that responses to GGCs can bypass deliberative and participatory contexts when the GGCs which constitute ‘wicked problems’ are ‘misframed as ‘tame problems’ that appeal to the expertise of dominant actors and their responses as solutions’ (Ludwig et al. 2022, 8). These cases show that there is a pressing need to address RI’s lack of critical reflection on how inclusion and participation can and should be integrated in practice.

Regarding the concept of deliberation, Michiel Van Oudheusden argues that ‘RI proponents have little or nothing to say about the politics and power that play out in, and through, deliberative governance processes’ (van Oudheusden 2014, 73). Using the TA project ‘NanoSoc’, which Van Oudheusden takes to be a practical implementation of RI in line with von Schomberg’s working definition, two things are highlighted: (1) the operative assumption that ‘involvement of more actors and issues in science would lead to better science policy and enhance scientific quality’ while at the same time (2) demonstrating an uninformed understanding of how deliberation actually works in practice as ‘cooperation and collaboration were taken for granted’ (van Oudheusden 2014, 77). The latter being what he regards as a ‘general lack of consideration for the politics in deliberation’ (van Oudheusden 2014, 77). In relation to these concerns, he further highlights the political nature and bias behind the choice of who to include in the deliberative context, thereby stressing how RI ignores questions related to the politics of deliberation. By doing this, he highlights the lack of critical reflection on RI’s own normative principles and ambiguity as to how they can actually be enforced in practice. This is one important sense in which RI can be said to deal with inclusion, deliberation, and politics only in a narrow sense.

In a similar vein, ‘[p]articipatory practices have become highly standardized and have been deployed as a technocratic repertoire of deliberative democracy that is propagated by a new class of democracy-for-innovation experts and mainstreamed across places and domains’ (Frahm, Doezema, and Pfotenhauer 2022, 180). Standardizing participatory practices may lead to a lack of sensitivity towards, for instance, contextual differences, diversity of subjects and communities, and the power dynamics involved in regulating and legitimizing such practices. This can undermine the very purpose and effectiveness of such practices.

The gap between RI’s conceptual and political shortcomings on the one hand and institutional interest in applying this framework on the other should not go unnoticed. For example, the lack of RI’s political and participatory dimension is sometimes seen as a merit rather than a limitation. This is the case for instance in a report released by the Stockholm International Peace Research Institute (SIPRI) which argues that RRI can be a useful approach to technology governance in mitigating risks in the context of the military use of artificial intelligence (AI). The report explains that its interest in the European RRI framework is due in part to its potential to bypass political challenges. It should be noted that SIPRI’s operative conception of politics is limited to the role of states and governments in the strategic high-stakes game that takes on the guise of ‘governance’. Politics in this context is therefore generally viewed as negative and a hindrance to efficient and responsible governance, but which could nonetheless potentially be bypassed by RRI. Furthermore, the authors of the report take RRI to be useful due to its aim ‘to involve all relevant stakeholders, particularly academia and industry, which have the technical understanding of the risks that may result from the development, diffusion and military use of AI technology’ (Boulanin, Brockmann, and Richards 2020, vii). This demonstrates how the operative conceptions of ‘inclusion’ and ‘participation’, in a suggested RRI uptake, struggles to go beyond an (expert) multidisciplinary approach and to incorporate a broader political dimension. As my argument progresses, we will see why this is the case, i.e. we will see which arguments are given to legitimize the exclusion of the public in deliberative and participatory practices.

Three studies on the political and critical deficits in RI

Public inclusion in discussions and negotiations that seek to identify societal problems, goals, and solutions is an important dimension of responsible innovation management. Or, to put it differently, including those who are impacted by the decisions made in the context of innovation governance is an important part of what it means to be responsible. Yet the concept of ‘inclusion’ is broad and its incorporation as a condition for responsible innovation management can be complex. For instance, including various voices in a debate does not directly imply that an agreement, which satisfies all parties involved, will be easily reached, if reached at all. In fact, assuming deliberation is free of strife is itself ideological, as such an assumption naively presupposes an interest shared equally by all (e.g. by the CEO of a company, the government, the shopkeeper, and the factory worker).

Furthermore, the relation between public inclusion and action in RI contexts is not always clear. How are the formulations of interests, concerns, and values of societal subjects impacted by RI frameworks actually negotiated? How are they decided upon and how are they integrated into the further development of the framework? Moreover, how do those formulations of societal interests, concerns, and values impact action (responsiveness)? In other words, how do the outcomes of deliberative processes impact action? How and to what extent is the public involved in the process dimension of responsiveness of the RRI framework? Or to pose the same question normatively: to what extent should the public be involved in the process dimension of responsiveness?

In what follows I will outline the results of three studies that identify different aspects of RI’s political deficits. The first study identifies how society, and specifically the idea of ‘societal participation’, is instrumentalized in RI frameworks to satisfy the goal of increasing competitiveness in the global economic arena. The second study shows that even where RI frameworks implement deliberative processes (to satisfy the ‘inclusion’-criteria), those processes have political elements that are often not considered, resulting in e.g. some (group) interests or ideologies being prioritized over others. The third study shows how participation is bypassed in responses to Global and Grand Challenges (GGCs) which require innovative solutions.

A new deficit logic: effects of the participatory turn in the RRI framework

The first study points out how the European Commission’s identification of societal participation as a ‘key’ for RI has led to ‘a new deficit logic’, namely a ‘democratic deficit of innovation’ (Frahm, Doezema, and Pfotenhauer 2022).³ This new deficit logic increased the perceived relevance of society, making it a key player in creating (good) innovations that can compete in the global liberal-market order. The authors of this descriptive and empirical study are critical of this ‘troublesome logic’ as it has re-packaged the relationship between technology and society; the new view is ‘an emerging and standardized view in policy-making that presents the integration of society as a ‘fix’ to problems with innovation policy and its contribution to global economic growth’ (Frahm, Doezema, and Pfotenhauer 2022, 203). The authors assert that ‘in practice these new politics of deficit construction get integrated with efforts to further advance, rather than break with, the innovative imperative and concurrent beliefs in technological fixes’ (Frahm, Doezema, and Pfotenhauer 2022, 177). In other words, inherent in the European Commission’s RI framework are the operative ideas that (1) the lack of innovation in a society or institution is seen as negative, (2) technological innovation is imperative to solving problems, and (3) societal participation is key to streamlining innovation governance such that innovation can bring about societally desirable outcomes. The identification of societal participation as pivotal to progress in the field of innovation and problem-solving has led to ‘a shift of authority over the coproduction of innovation and democracy from the national to the global arena’ (Frahm, Doezema, and Pfotenhauer 2022, 178). This in turn provokes critical questions: How does the Responsible Innovation framework, as one that emphasizes the role of societal participation in the management of innovation processes, gain legitimacy in the institutional context of the European Commission? (Frahm, Doezema, and Pfotenhauer 2022) How can societal participation be implemented in practice without itself becoming the very thing it is trying to avoid (e.g. an overly standardized practice that serves particular agendas and gives rise to a new form of ‘expert’)? (Frahm, Doezema, and Pfotenhauer 2022) I press further here by adding: How can RI prevent that ‘inclusion’ or ‘participation’ translates, in practice, into experts, companies, and institutions persuading society into finding particular innovations (as responses to specific problems) acceptable? Ultimately this last question assumes that RI has a responsibility to ensure that societal actors do not become the weakest ‘stakeholder’, but rather to ensure that they not only have a seat at the table, but that they are actually seen and heard.

Frahm et al.’s study of policy documents released by the European Commission shows that it claims ‘a considerable power to identify national and organizational deficits in the ways publics are constructed, educated, and induced to participate in the governance of innovation’, simply by virtue of ‘[g]aining authority over how RRI is defined’ (Frahm, Doezema, and Pfotenhauer 2022, 197) and the normative framework it creates. This translates into the significant power to shape public engagement and opens up the possibility of misusing this power to exclude certain perspectives in the definition and implementation of RRI. The study further shows that the European Commission implements micro-interventions to ‘subtly position [itself] as the global arbiter of what responsibility in research and innovation entails in practice, and by association right modes of reason and democratic participation’ (Frahm, Doezema, and Pfotenhauer 2022, 199). This is no small claim. Gaining authority over the definition of ‘responsibility’ in RRI and by extension setting up the normative parameters for related concepts such as participation and democracy is a political matter. While the authors do not expound on the implications of this, I take it as (1) an example of the importance of integrating a critical method, including immanent critique and (2) the need for more critical consideration of the place of participation in the RRI framework, i.e. when is (public) engagement and intervention considered legitimate from the perspective of the European Commission? From the perspective of RRI, part of what it means to innovate responsibly is to include various stakeholders (including the public) in its deliberative processes. However, little effort has gone into questioning whether the public is (and should) be included and allowed to participate in the design of the RRI framework itself. Including the public in the design of the governance framework itself could prevent subtle attempts to bypass participation in implementations of R(R)I that result from the way in which the ‘right modes’ of reasoning and participation are framed.

It is important to engage in critical reflection on the European Commission’s normative framing of concepts in the RI context given that this framing reflects interests on a micro level and operative rationality on a macro level. For example, the European Commission endorses (public) participation because it is understood to be instrumental in improving economic competitiveness in the global context – demonstrating the kind of instrumental rationality at play. Society is made relevant because ‘[d]emocratic deficits in national innovation policy are presented as potentially impeding global economic growth and competitiveness’ (Frahm, Doezema, and Pfotenhauer 2022, 201). However, the instrumentalization of society to serve the global liberal-market order is not unproblematic, especially alongside powerful institutions that have developed the authority to determine a normative framework of what counts as responsible methods, and hence solutions, to pressing societal challenges. This questions the agency and freedom with which societal actors can actually contribute and participate in the discussion. While the authors of the study mentioned above point out that the ‘epistemic authority over democratic matters with regard to innovation (…) has considerable political implications’ (Frahm et al., 203), it is not clear how this issue can be addressed moving forward. By engaging with critical theory, I will indicate a fruitful direction for Responsible Innovation to take in relation to the European Commission.

Picture perfect democracy? The difficulties of deliberation

The instrumentalization of society to achieve innovation and economic goals, as well as the authority established and reproduced by the European Commission, is just one illustration of the political challenges facing its RRI framework on a macro level. There are also problems with the way in which the RRI framework and its uptakes have tried to fulfill the criterion of ‘inclusion’ or ‘participation’ both conceptually and in practice.

European RI frameworks have been criticized for ignoring ‘questions about the politics in deliberation … as well as the politics of deliberation’ (van Oudheusden 2014, 68). Regarding the politics in deliberation, this criticism asserts that RI frameworks display a naïve understanding of the politics and power dynamics at work in deliberative contexts, instead assuming that deliberation entails a potential consensus, bracketing the possibility of conflict provoked by competing interests and values. RI proponents ‘emphasize the importance of talk, deliberative argumentation, and due procedure without attending to questions of power, ends, and authority that play out in, and through RI processes’ (van Oudheusden 2014, 69). As a result, it is unclear how RI frameworks aim to deal with the practical reality of deliberation. The criticism shows that there is a gap between the concept of (democratic) deliberation and how it plays out in practice.

The criticism of RI’s conceptual negligence of the politics of deliberation highlights that there are political biases and choices made in the construction of a deliberative context. This criticism suggests that there are operative conceptions of democracy and deliberation that determine the conditions for legitimacy. However, one can push this criticism further and argue that even the operative conceptions of democracy do not live up to the ‘democratic ideal’. While inclusion is at the heart of the ‘democratic ideal’, it is difficult to realize institutionally and therefore usually leads to democracy with shortcuts (Lafont 2019). While there are different institutional manifestations and theoretical examples of democracy with shortcuts – e.g. deep pluralism, epistocracy, and lottocracy – they all have in common that at a certain point, those subjected to the institutions in question are expected to blindly defer ‘to the political decisions of others’ (Lafont 2019, 356). Institutions can claim to prioritize inclusion – in the name of democracy – but when the wills and opinions of those subjected to the institutions are bypassed in practice, the (institutionally mediated) deliberative process is rendered obsolete. In my view, this again emphasizes the importance of critical methods such as immanent critique to (1) reflect on the operative concepts (e.g. inclusion, democracy) and assumptions (e.g. of deliberative practices, political legitimacy) held by RI frameworks and to (2) critically examine if these RI frameworks can live up to the standards they set out for themselves given these operative concepts and assumptions.

Bypassing participation: a critique of how global and grand challenges are negotiated

Although there is increased sensitivity to include the public in RI processes, the third study reveals that public participation can still be circumvented in practice by means of a rhetorical strategy that labels problems that are actually ‘wicked’ as ‘tame’, downplaying the complexity of the given problem and hence the need for public participation in finding a responsible solution. More specifically, the third study shows that despite the prominence of ‘Global and Grand Challenges’ (GGCs) in RI discourses, little attention has been given ‘to the question of how responses to GGCs are negotiated’ (Ludwig et al. 2022, 7).

According to the study, responses to GGCs can take the form of a negotiation strategy or a solution strategy. The former stays close to society by implementing forms of participatory governance and public engagement while the latter appeals specifically to ‘the unique expertise of scientists and innovators’ (Ludwig et al. 2022, 10). The results of the study show how ‘GGCs often set the stage for legitimizing dominant responses as solutions rather than encouraging reflexivity about heterogenous perspectives on contested social-environmental problems’ (Ludwig et al. 2022, 7). Thus, participatory practices can be bypassed when ‘wicked problems’ – which in fact characterize GGCs – are ‘misframed as ‘tame problems’’ (Ludwig et al. 2022, 8). This is an issue because the formulations and solutions to ‘wicked problems’ are disputed and therefore should be negotiated. Instead, when GGCs are incorrectly framed as ‘tame problems’ the solutions ‘appeal to the expertise of dominant actors’ (Ludwig et al. 2022, 8) and therefore circumvent the space in which public participation practices can take place. This insight is highly pertinent as it suggests that the rhetoric of GGCs is complicit in limiting participation in responsible innovation governance, thereby problematizing the possibility of a better relationship between society and science, innovation, and technology.

The study also showed that research organizations such as Wageningen University and Research (WUR) have begun to adopt ‘GGCs and responsibility as organizing concepts’ (Ludwig et al. 2022, 12). In the last decade, the uptake of these concepts has shaped its identity as a research institution, in turn affecting its practice (Ludwig et al. 2022, 13) – in other words, what an institution thinks affects what it does. The authors argue that ‘[r]esponsibility at the substantive level of solving GGCs does not imply responsibility at the procedural level of negotiating them’ (Ludwig et al. 2022, 14), which they exemplify by showing the results of a document analysis of WUR’s Annual Reports and Strategic Plans. The results demonstrated how support for the ‘solution strategy’ was implicit in these documents through references that appealed primarily to ‘the university [WUR], its research/researchers, or its education/students’ and more broadly ‘the ‘triple helix’ of universities, government, and business’ (Ludwig et al. 2022, 14) as actors for addressing GGCs. Little reference was made to either NGOs (2% of total references) or local stakeholders/citizens (2% of total references), less than 1% of references appealed to society, and no reference was made to marginalized stakeholders.

This trend of appealing to a solution rather than a negotiation strategy can be identified in other contexts as well, such as research institutes that seek to inform and promote RI approaches and practices to policymakers, researchers, and the broader interested public. The example we analyze here is the report mentioned earlier on ‘Responsible Artificial Intelligence Research and Innovation for International Peace and Security’ put forth by the Stockholm International Peace Research Institute (SIPRI) in November 2020. The report refutes the belief that the ‘development, diffusion and adoption of military and dual-use application of AI’ is determined, but argues rather that ‘it is a choice, one that must be made with due mitigation of risks’ (Boulanin, Brockmann, and Richards 2020, vii). They believe RRI ‘should lead to decisions in the innovation and commercialization processes that can help to prevent, or pre-emptively mitigate, risks associated with the development, diffusion and military use of AI’ (Boulanin, Brockmann, and Richards 2020, viii). They therefore take RRI to be a useful technological governance framework. Yet the way they argue for RRI exposes how the ‘solution strategy’ is implicit in the way they conceptualize it. They take the primary relevant stakeholders in this context to be academia, industry, engineering, and science because they ‘have the technical understanding of the risks that may result from the development, diffusion and military use of AI technology’ (Boulanin, Brockmann, and Richards 2020, vii, emphasis mine).

The researchers believe RRI can bypass the political challenges that slow down the process of defining risks and implementing policies for military uses of AI. They are concerned that advancements in AI technology will outpace progress in arms control (Boulanin, Brockmann, and Richards 2020, 9). They therefore understand ‘an inclusive and multi-stakeholder approach’ as involving a ‘diverse array of actors in research, academia, the private sector and government’ (Boulanin, Brockmann, and Richards 2020, 12) – or the triple helix referred to earlier. The pressing nature of the issue (i.e. AI technology outpacing arms control) appears to justify the exclusion of certain relevant stakeholders. However, limiting the concept of an inclusive and multi-stakeholder approach to the groups listed above risks neglecting the importance of public participation and by extension alienating the public from discussions regarding peaceful (co-)existence and security. By not explicitly considering the public as a key player in a multi-stakeholder approach, the authors imply that the public does not have a role or stake (or has a weak role) in the responsible governance of military uses of AI. This approach to RRI is an example in which political and public involvement are bypassed in favor of an expert-based approach. It assumes that the knowledge of experts is both necessary and sufficient with regard to the responsible governance of military uses of AI. This assumption prioritizes reaching a consensus quickly over considering diverse societal perspectives.

This brings us back to the conclusion inferred by the authors of the third study. They inferred that the wicked character of GGCs are often rebranded as tame problems to bypass negotiation and participatory practices. This seems to stem from the tension between the urgency to intervene in GGCs and the long road of negotiation and participation even though ‘it is precisely this move towards intervention without negotiation that reinforces dominant perspectives and that thereby often leads to responses that deepen inequality between stakeholders’ (Ludwig et al. 2022, 22). This reflex to bypass participation has also been detected in the report released by SIPRI. Ludwig et al. conclude that there is a ‘need for a critically reflexive debate about navigating heterogeneity that meaningfully gives way to diverse innovation futures’ and includes ‘a substantial rather than superficial appeal to negotiation in which heterogeneity is actually shaping research and innovation rather than merely legitimizing them’ (Ludwig et al. 2022, 22). This is a good starting point but what does such a critically reflective debate look like? How can we ensure that heterogeneity is actually shaping research and innovation? The authors seem to presuppose that addressing these issues constitutes a responsible attitude towards research and innovation processes. My suggestion would be to integrate critical and participatory practices into RI frameworks, enabling immanent challenges to RI’s operative rationality.

To summarize, it is not possible to ignore the political dimension of technological innovation as the question of what to innovate and how to innovate is also a political matter. Critique has a central role in these questions. A framework of ‘Responsible Innovation’ therefore needs to include political and critical dimensions in order to gain legitimacy on its own terms. It is thus highly problematic that the RRI framework put forth by the European Commission suffers from political deficits, consequently impacting its (institutional) uptake. The rather well-received claim that RI frameworks suffer from political deficits has invited reflections on how to conceptualize this deficit and how to address it, not only conceptually, but in practice.

The three main critical studies examined thus far share a common theme, namely a concern for the power relations implicit in RI frameworks and uptakes. They have shown for instance how these power relations play out on the macro level between individuals and communities on the one hand and institutions (such as the European Commission) and its experts on the other. We have also seen how they play out on the more micro level, within the context of deliberative frameworks, for instance. However, central to an analysis of power is an understanding of the social relations and social structures in a given context, as these are complicit in shaping and determining the ways in which power is acquired, exercised, and distributed in societal and political contexts. Noticeably, a critical analysis of social relations and structures is lacking from the critical literature that addresses the political deficits of RI. This makes an approach that incorporates critical theory well-suited to analyze and understand the ways in which power and social relations and structures (negatively) impact RI frameworks and uptakes.

On with critique! On the responsibility for emancipation

As we have seen, the European Commission’s RRI framework takes responsible governance to include a wide variety of societal actors (with diverse perspectives, needs, and values) in the innovation process. RRI brands itself as socially responsible, taking seriously the relationship between society and science and technology. Yet the uptake and implementation of this form of societal inclusion has proved to be challenging. As the studies outlined above have demonstrated, the RRI framework suffers from political shortcomings, both conceptually and in practice, that problematize its own inclusive and participatory agenda. I argue that engaging with critical theory can be immensely valuable for the further development of a politicized and responsible innovation framework. This of course (re-)opens the important question: What is meant with ‘responsibility’ in the RI context? While the answer to this question has many dimensions, my focus here is on the role and importance RRI specifically attributes to inclusion and public participation. Specifically, I suggest transforming the RRI framework by incorporating the emancipatory ambitions that have motivated critical theorists since the 1930s and making use of participatory and critical practices to realize them.

In this section, I will first explain why critical theory's general emancipatory ambitions are relevant for the RRI framework. Then, I will focus on the work of Andrew Feenberg, who has long advocated for the democratization of technology and whose arguments are highly pertinent for a politicized and responsible innovation framework. The argument that technological innovation has emancipatory potentials is not foreign – e.g. in the context of combating the environmental crisis, the argument has been made that new and emerging technologies have the ‘emancipatory potential […] to inaugurate the post-Anthropocene’ (Blok 2022, 17) – yet very little is said explaining what is meant with emancipatory potential, how it can be understood on the social level, and how it can or should be achieved. This is especially important for the RRI framework as it claims to be socially responsible or responsive to society. My goal here is to invite critical reflection on the potential of technological innovation to advance social emancipation – starting at the local level – and the role of RI frameworks to facilitate it.

While there are important differences between contemporary critical theorists and those of the Frankfurt School’s first-generation, the original ‘emancipatory ambitions’ of the school of thought have endured (Deutscher and Lafont 2017, xiv). In the words of Raymond Geuss: ‘Critical theories aim at emancipation and enlightenment, at making agents aware of hidden coercion, thereby freeing them from that coercion and putting them in a position to determine where their true interests lie’ (Geuss 1981, 55). This presupposes that agents can have ‘false’ interests or needs, that appear as true interests or needs and further that exposing this coercion can lead to greater emancipation. In my view, the normative orientation in critical theory to aid in the identification and de-naturalization of ‘hegemonic patterns of interpretations’ (norms) (Honneth 2018, 919) and the exposure of oppressive and coercive social relations and structures, harbors an implicit responsibility, not only to understand society but to change it with the ambition of greater social emancipation. Identifying and (immanently) critiquing the social norms, relations, and structures that create, exercise, and perpetuate forms of oppression and inequality with the aim of greater emancipation is what normatively characterizes what I consider the responsibility for emancipation. Together with social theory, critical theory can aid in highlighting the agency and critical practices of everyday social actors and identify the barriers to critique implicit in participatory frameworks. As will be shown in the following passages, while societal inclusion and the creation of spaces for critical and participatory practices are crucial for responsible innovation frameworks, critically reflecting on the biases implicit in them is equally important.

Feenberg is an important dialogical partner in this endeavor. Not only is he a critical theorist – inspired greatly by his teacher Herbert Marcuse – but he has made significant contributions to the philosophy of technology and STS with his critique of technology and long advocacy for democratizing technology. His work is very relevant for RI frameworks, especially since the concept of ‘innovation’ has garnered its specifically commercial and technological connotations (Godin 2015; von Schomberg and Blok 2019). Yet it is also by virtue of RI’s own status as a technology – in the sense that it provides a systematic approach to the management of innovation processes as well as criteria (or mechanisms) to monitor and evaluate that management – that Feenberg’s work gains increasing relevance.

Feenberg has opposed the idea of technological determinism and prominently promoted the democratization of technology since the 1990s. He has written about the importance of democratic intervention, using the example of the environmental movement of the 1990s.⁴ Environmentalists at the time demanded alternative energy and regulation of existing technologies, which was met with corporate and political pushback (Feenberg 2015, 500). These ‘[d]emocratic interventions into technology … challenge the horizon of rationality under which technology is … designed’ (Feenberg 1999, 14). This in turn can lead to the demands being translated into new technical designs or regulations (Feenberg 2015, 501). The democratic interventions of activists should not be excluded from our discussion on how to achieve responsible innovation. According to Feenberg, micropolitical activism is vital to a technological society, despite its pitfalls,⁵ because it ‘limits the autonomy of experts and forces them to redesign the worlds they create to represent a wider range of interests.’

Activism⁶ is a clear form of critique, as activists make claims and undertake actions – not limited to discourse – to express discontent and achieve aims. It typically falls outside the parameters of institutional participatory frameworks, yet, as has been argued earlier, these frameworks are biased by design. Recent research has demonstrated how activists can apply pressure to firms by making certain claims about their practices, resulting in the adoption of more responsible practices even in cases where the firms in question are initially resistant to adopting them (Waldron et al., [2019] 2022). This study focuses exclusively on the external pressures exerted by activists onto firms and not on the role of institutions in applying this pressure.

Although more empirical research is needed to examine the role of activism in the responsible shaping of technological innovation processes, it is clear that there are barriers to critique owing in part to the implicit biases structuring participatory RI frameworks. These barriers in turn may hinder critical efforts (including activism) to contribute to RI discourses.

In the late 1990s, Feenberg identified the fundamental problem of democracy as ‘quite simply the survival of agency in a technocratic universe’ (Feenberg 1999, 4). Agency translates here into the capacity of individuals and groups to insert themselves and transform the social and political structures in which they exist. Feenberg argues:

All forms of public activity and participation should be sanctioned as democratic as long as they respect the rights of others. Democracy therefore includes citizens’ attempts to enhance participation and agency by reforming the procedures of government, business, education, and other social spheres. (Feenberg 1999, 4)

The enhancement of agency is therefore included in what is considered legitimate ‘democratic participation’. Impediments to agency or participation, such as technocracy, threatens democracy itself as it ‘offers persuasive arguments and alibis for passivity’ (Feenberg 1999, 4) – or to put it differently, it depoliticizes individuals and communities.

Technocracy armors itself against public pressures, sacrifices values, and ignores needs incompatible with its own reproduction and the perpetuation of its technical traditions. (Feenberg 2006, 181)

Agency and democracy are central to Feenberg’s technological optimism. I share his view that democratic interventions (or public participation) can succeed ‘in orienting innovation toward new types of designs’ (Feenberg 2006, 201). I take it to be critical for those subject to the outcomes of technological innovation processes to be incorporated in questions concerning its design.

The need for and impact of technological innovations will vary depending on a given context. Individuals and groups within those contexts have unique and situated experiences, perspectives, and values that need to be heard and considered. It seems vital to responsible innovation practices and governance frameworks to appreciate the given social context and the pre-existing social relations in which innovative solutions will be introduced. This is because rationality is itself situated – e.g. planned obsolescence is rational in the capitalist context but irrational abstracted from that context. The integration of critical and participatory practices in innovation processes has the potential to challenge operative rationality. This in turn increases the possibility of attaining greater forms of emancipation.

The existence of different kinds of rationalities presupposes rational alternatives. The alternative chosen in a particular technological design carries with it a bias. Before we can continue, Feenberg’s distinction between ‘substantive bias’ and ‘formal bias’ needs to be clarified, a distinction he borrows from Max Weber’s theory of rationality (Feenberg [1991] 2002, 81). Substantive bias refers to the more everyday connotation of ‘prejudice and discrimination’, while formal bias ‘concerns the rational form of artifacts and social arrangements’ (Feenberg [2020] 2022, 30, emphasis mine). Substantive bias ‘rests on a content of belief as such’ (Feenberg 2017, 22). For instance, a claim or belief can be substantively biased if it is prejudiced or discriminatory. While a critique of substantive bias ‘still plays an important emancipatory role’ (Feenberg 2017, 22), such a critique often implicitly contrasts bias with reason (with the latter assumed to be neutral and universal). Appealing to reason to critique bias, for instance, ‘grants the neutrality and universality of institutions [broadly conceived] that claim a rational foundation’ (Feenberg 2017, 22). Following the lineage of Jean-Jacques Rousseau and Karl Marx, Feenberg does not assume that bias and rationality are necessarily opposed. Rational, well-designed systems or technical devices can exist that are nonetheless biased (Feenberg 2010, 69). These systems or technical devices can be biased regardless of the presence of substantive bias on the side of the user of the system or device. The concept of formal bias is thus needed to critique rational systems. This is because even where substantive bias can be avoided, formal bias can still be at play (Feenberg 2010, 69).

Formal bias hides in aspects of rational systems that only become visible in the light of historical and contextual analysis. It is not a matter of prejudice based on pseudo-facts or narrative myths; rather, the design of the system objectifies the discriminatory principle. (Feenberg 2017, 24)

The role of rationality is important here and we can see the influence of Hebert Marcuse and other members of the Frankfurt School. In the words of Marcuse: ‘one of the most vexing aspects of advanced industrial civilization [is] the rational character of its irrationality’ (Marcuse [1964] 1968, 9), and further that ‘technological controls appear to be the very embodiment of Reason for the benefit of all social groups and interests – to such an extent that all contradiction seems irrational and all counteraction impossible’ (Marcuse [1964] 1968, 9) – recall here the reference to how technocracy depoliticizes individuals and communities.

Similarly to Marcuse, Feenberg often refers to capitalism as a social structure, to show that while it is technically rational, its formal bias informs its technical decisions. A governance framework, such as RI, for instance, operates not only within an institution that has particular interests (e.g. the European Commission), but within a larger social structure (e.g. capitalism, which mediates social relations). Both ‘social’ and ‘structure’ are important concepts. ‘Structure’ limits and conditions the possibilities of action, while the adjective ‘social’ points to the role of human agency in creating and reproducing those structures (along with the social and power relations within them). We can also explain this more generally: through their activities, humans create and build the world with its structures and institutions. This world also conditions humans, it creates dimensions of rationality in which our activities ‘make sense’. In this sense, social structures stabilize human life against the contingency of nature (Arendt [1958] 1998). It is also from within these structures that the potentiality to change them resides, in the form of agency. While the motivation to challenge and transform operative structures and the formal bias that determines them is usually provoked by the perception or experience of dissatisfaction or injustice.

Critical and participatory practices are important to be able to identify ‘formally biased technical codes’ (Feenberg 2010, 69) that determine technical decisions and designs and by extension, impacts the ability of the user to maneuver within the technical system. A technical code ‘is a criterion that selects between alternative feasible [technically workable] technical designs in terms of a social goal and realizes that goal in design’ (Feenberg 2010, 68). Technical codes are therefore a realized ‘interest or ideology in a technically coherent solution to a problem’ (Feenberg 2010, 68). The deliberate choice – between materially and technically possible alternatives – is tainted by its formal bias; ‘[t]he choice itself results from the play of dominant interests’ (Marcuse [1964] 1968, xvi), especially where critical and participatory practices are bypassed. Marcuse argues that the ‘choice between historical alternatives’ is ‘one ‘project’ of realization among others. But once the project has become operative in the basic institutions and relations, it tends to become exclusive and to determine the development of the society as a whole’ (Marcuse [1964] 1968, xvi). This resonates, on a more micro level, with the criticism of the first study I examined (authored by Frahm et al.), which expressed concern regarding the power and autonomy gained by the European Commission as it positions itself as a global arbiter of what counts as legitimate forms of reason and democratic participation in the RRI context. By extension, the concern is reinforced that there are exclusionary mechanisms implicit in the way technocratic institutions (such as the European Commission) frame participatory practices. This latter point is also reflected in the third study I examined (authored by Ludwig et. al.). This study demonstrated that ‘negotiation strategies’ (which have participatory dimensions) are considered important when problems are characterized as ‘wicked’ rather than ‘tame’. This in turn enabled those in a position to classify problems to ‘mischaracterize’ wicked problems as tame ones in order to bypass the need for negotiation. The calculable criterion of efficiency lurks in the background here because the quicker a solution is found the better. This efficiency provokes us to recall the critiques of instrumental rationality formulated by several first-generation Frankfurt School theorists. Yet as Feenberg reminds us: ‘[t]echnology is ‘underdetermined’ by the criterion of efficiency and responsive to the various particular interests and ideologies that select among [several viable design options]’ (Feenberg 2010, 67).

Participatory practices are integral to critical practices because they draw attention to possible (innovation) alternatives and serve as a reminder that we live in a world with heterogeneous perspectives and experiences, in turn challenging hegemonic social norms and structures and the idea of an ‘ultimate’ rationality that deters counteraction. Bypassing participation determines innovation processes in a particular way. In Feenberg’s words:

The exclusion of workers and communities from participation in technical decisions means that some considerations relevant to human wellbeing will be overlooked. […], that exclusion also biases technical designs and the technical disciplines themselves. (Feenberg [2020] 2022, 26)

In my view, this is why RI frameworks should be oriented toward social emancipation through participatory and critical practices. This entails that there need to be spaces where these practices can take place such that formally biased technical codes – which also create institutional barriers to critique – can be identified. RI frameworks should therefore be urged to take seriously local and situated forms of critique that arise when individuals and communities have a real seat at the table. It should also critically reflect on the social norms and structures that mediate social relations and bias technical designs. Recalling that RI frameworks are themselves technologies that are part of a social structure, they need to integrate the practice of immanent critique to be able to identify the formally biased technical codes that shape their design and practices. Given that there are many barriers to critique, immanent critique does not presuppose that everyday social actors are in a position to articulate their criticisms or social struggles (as a purely participatory or deliberative approach would presuppose). Immanent critique can therefore help identify the barriers to critical, public participation caused by e.g. the biased parameters delineating a technical framework such as a deliberative one. This extends to the more micro level as well, in line with the criticism of the second study I considered (authored by Van Oudheusden) which implied that RI frameworks need to critically reflect on their operative political concepts (e.g. deliberation, participation) in order to detect political and power dynamics and reduce ambiguity regarding how these concepts are used and should be applied.

Conclusion

This paper sought to show how the political and participatory deficits of R(R)I can be addressed by recourse to approaches inspired by critical theory and its emancipatory ambitions. I outlined the results of three studies, each of which highlighted different aspects of the political and participatory deficits of R(R)I. The common theme detected in all three studies was the concern for the power relations implicit in RI frameworks and uptakes. I have argued, in line with a critical theory approach, that an analysis of power requires an understanding of the hegemonic social norms, social relations, and social structures in a given context and that this is lacking in the critical reflections on current R(R)I frameworks and uptakes.

The work of Andrew Feenberg has provided a foundation to critically reflect on the role of bias in technical designs (of which RI frameworks are included). This poses the questions: Why is one technically rational alternative chosen over another? What norms and interests, as well as social relations and structures, are complicit in determining how technological innovation processes unfold? Should technological innovation be geared toward social emancipation (alleviating forms of social suffering)? Recognizing the agency and critical practices of everyday social actors is important in approaching these questions as those actors deal with the possible repercussions of innovation-solutions. While Feenberg has highlighted the role of democratic, public interventions (e.g. activism) into technical issues, more empirical research is needed to understand their role in influencing technological innovation processes over time. This includes identifying and addressing barriers to critical practices; this identification is also disclosive as it raises the question of why there are barriers in the first place. Reflecting on these topics can shed light on what is meant with responsible innovation and encourage efforts to highlight the responsibility for emancipation that has motivated critical theorists for decades.

Acknowledgements

I would like to express my deepest gratitude to Blake Scott and the peer reviewers for their helpful and constructive feedback. I would also like to extend a word of appreciation to the JRI Editorial team for their assistance.

Disclosure statement

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

Notes

1 The normative uncertainty brought about by modern technological developments has been a topic of concern for philosophers for decades. See in particular Hans Jonas’ concern in The Imperative of Responsibility ([1979] 1984).

2 In this article, I will use RI to denote the more general discussion of Responsible Innovation, and RRI to denote it’s specific uptake by the European Commission.

3 The article of Frahm et al. uses not only the European Commission as an instance of this new deficit logic, but also the Organization for Economic Cooperation and Development. However, for the purposes of this paper, the focus will only be on their treatment of the European Commission’s Responsible Innovation framework.

4 In his work Feenberg also references the student movements of the 1960s and the feminist movements of the 1970s as further examples of democratic interventions that intersect with technical issues (See Feenberg 2015, 501).

5 The main pitfalls Feenberg identifies include the possible “lack of long-term organization” and tendency to focus “on a single issue and sometimes a single location” (Feenberg 2015, 501)

6 The way activism is understood here is in line with Feenberg’s criteria of democratic public participation as respecting the rights of others in their attempts to enhance agency (See Feenberg 1999, 4).