Sage Journals: Discover world-class research

Abstract

Given the excessive power of Google and other large technology firms, transparency and accountability have turned into matters of great concern for organization scholars. So far, most studies adopt either a causal or critical perspective on the relationship between the two concepts. These perspectives are pitted against each other but share some basic assumptions – a fact which limits organization theory’s ability to fully grasp the management of (digital) visibilities. To address these limitations, we therefore propose a third, constructive perspective on these concepts. A constructive perspective turns transparency and accountability from analytic resources into topics of inquiry, allowing organization scholars to study how people in and around organizations put them to work and with what consequences. We introduce sites of ethical contestation as a new methodological strategy to conduct surprising and unintuitive empirical research from a constructive perspective.

Keywords

accountability ethics ethnomethodology open strategy transparency

Introduction

At its annual developer conference in May 2017, Google CEO Sundar Pichai announced the firm’s move from a ‘mobile first’ towards an ‘AI first’ strategy. While the rise of smartphones had prompted Google to radically redesign their existing products around the affordances of mobile computing, the growing capacities of intelligent technologies now urged the company once more to fundamentally revamp their products. While this strategic shift might have come as a surprise to some people in the audience, it was foreshadowed by Google’s growing interest in the transparency and accountability of AI technology. Only a few months prior to the CEO’s strategy presentation, Google co-founded and broadly promoted the non-profit organization Partnership on AI (PAI) as a platform to conduct and make openly available research on the ethics of AI and to develop best practices for the technology industry at large. Through this move, Google presented itself as a firm which strategically embraces a new and powerful technology but at the same time ensures greatest accountability through unprecedented transparency. However, despite this explicit commitment to transparency and accountability around their new strategic focus, public criticism regarding Google’s AI efforts has burgeoned ever since. Among other things, critics argued that Google would ‘outsource moral responsibility’ when developing AI for military use (BBC News, April 5, 2018), that it made ‘ethical mistakes’ when demonstrating the language proficiency of its AI assistant (Zerega, 2018), and that it raised ‘concerns about the ethics of data harvesting’ when revealing details about its advancements in facial-recognition AI (Wong, 2019).

The case of Google’s conflict-laden strategic shift towards AI illustrates a perplexing observation about organizations and their stakeholders in digitally networked environments: On the one hand, many of the most powerful organizations embrace transparency and accountability as their core values and standard business practices. New digital tools and technologies allow these organizations to share unprecedented amounts of data and code with their stakeholders. On the other hand, citizens, activists and policy makers use digital platforms to organize large-scale and enduring protests against these corporations, accusing them of opacity, unethical business practices, and a lack of accountability. The more organizations embrace and stakeholders demand transparency and accountability, the more difficult it becomes to understand how these concepts relate to each other. Does the greater accountability, which Google claims to create through its engagement for transparent AI, accelerate or rather hinder the firm’s strategic shift? Does transparency of a technology such as AI make an organization which uses this technology more accountable? Do critics bemoan a lack of accountability due to superficial transparency or based on what they learned from gazing into radically transparent organizations? In this essay, we argue that this sense of confusion is nurtured by organization theories that are insufficiently equipped to fully grasp the role of transparency and accountability in everyday organizing.

So far, most organizational scholarship has studied transparency and accountability from either a causal or a critical perspective (for recent reviews see Albu & Flyverbom, 2019; Bernstein, 2017). Studies that adopt a causal perspective believe that broader and deeper transparency allows stakeholders to more effectively enforce norms of ethical business practice (e.g. Cucciniello, Porumbescu, & Grimmelikhuijsen, 2017; Schnackenberg & Tomlinson, 2016; Whittington & Yakis-Douglas, 2020, this issue). In contrast, research that takes a critical perspective argues that the relation between the two concepts is more complex and that any form of transparency can only ever create incomplete accountability (e.g. Hansen & Flyverbom, 2015; Hansen & Weiskopf, 2019; Heimstädt, 2017; Ringel, 2019). Both perspectives have contributed significantly to our understanding of ‘visibility management’ (Flyverbom, 2020) in and around organizations. However, although they are pitted against each other they share some basic assumptions, namely that they rely on transparency and accountability as predefined researcher-analyst categories which exist in a theoretical relationship to each other. Having two seemingly antagonistic perspectives which share a similar underlying assumption limits organization theory’s ability to fully unpack the puzzling simultaneity of transparency, accountability and ethically controversial business practices found in cases such as Google’s AI strategy.

In this essay, we therefore propose a constructive perspective as a new vista for analysing transparency and accountability. To shift from a causal or critical into a constructive perspective, organization scholars can turn transparency and accountability from analytic resources for studying empirical phenomena into topics of inquiry and phenomena in their own right. Turning resources into topics allows researchers to analyse how people in and around organizations – practitioner-analysts – mobilize these topics in their everyday activities to create social order (Pollner, 2012; Zimmerman & Pollner, 1970). Studying analytic resources as topics of inquiry means focusing on how people in and around organizations claim, contest and configure – i.e. construct – these concepts, in which situations and with what consequences.

The constructive perspective comes with the need for new and creative methodological strategies. We therefore propose that organization scholars should study transparency and accountability through what we call sites of ethical contestation. We define these sites as spaces in which new and ambiguous matters of transparency and accountability are claimed, contested and configured. Their key characteristic is a lack of ontological security regarding these concepts, which requires people to explicate and justify their conflicting definitions of transparency and accountability. While sites of ethical contestation can usually be found at the fringes of organizations, their form can vary greatly ranging from ordinary shareholder meetings to more unusual situations such as meetings between whistleblowers and journalists. To examine this wide variety of sites, we suggest that organization scholars use, adapt and combine at least three research tactics: tracing provisional practices, comparing staged performances and researching the researchers. Turning towards a constructive perspective and this new methodological vista, organization scholars can move beyond research on the abstract relation between transparency and accountability and start exploring how people use transparency and accountability as tools for organizing work in digitally networked environments.

Causal Perspective: Transparency Creates Accountability

Google’s commitment to transparency and accountability on AI can be interpreted as an attempt to evade ‘hard’ governmental regulation by embracing ‘soft’ self-regulation. In organization studies this interpretation resonates with what we call the causal perspective on transparency and accountability. Research that takes a causal perspective is generally optimistic that more transparency can lead to more accountable organizations. If organizations disclose information about their inner workings, the story goes, they have ‘little alternative but to adjust their deeds to match societal expectations’ (Zyglidopoulos & Fleming, 2011, p. 692). These adaptations can either take place in an anticipatory fashion and before the information is disclosed, or as a response to negative feedback from their audiences. Working from the firm assumption that transparency can create accountability, research from a causal perspective asks how this result can best be achieved. Answers to this question oftentimes revolve around the ‘right’ calibration of transparency, i.e. what type and what amount of information is disclosed, to which audiences, at what time, speed and frequency. To many practitioners and organization scholars alike, the causal perspective sounds immediately intuitive, not at least because it is firmly anchored in popular and political culture. For example, proponents of the causal perspective oftentimes refer to the political slogan that ‘sunlight is the best disinfectant’, initially coined by US Supreme Court justice Louis Brandeis (1914) and since then mobilized by generations of activists, journalists and policy-makers.

The causal perspective and its underlying idea that ‘seeing means knowing’ has been widely used in domains of organizational research such as public management (e.g. Cucciniello et al., 2017) or corporate social responsibility (e.g. Haack, Martignoni, & Schoeneborn, 2020). More recently it has also been applied to one of the field’s core domains: strategy-making. As firms increasingly adopt more transparent forms of strategy-making (Whittington, 2019; Whittington, Cailluet, & Yakis-Douglas, 2011), researchers begin to examine the outcomes of these changes. While some of the existing work focuses on the effects of greater transparency in strategy-making on organizations’ financial performance (e.g. Whittington, Yakis-Douglas, & Ahn, 2016; Yakis-Douglas, Angwin, Ahn, & Meadows, 2017), others start to explore the consequences of transparent strategy-making for the relation between organizations and stakeholders more broadly (e.g. Dobusch, Dobusch, & Müller-Seitz, 2019; Gegenhuber & Dobusch, 2017). Exemplary for the latter, Whittington and Yakis-Douglas (2020) theorize that greater transparency around strategy-making can lead to more accountable organizations. When organizations become more transparent around strategic issues through ‘managed’ or ‘unmanaged practices’ of openness, they argue, ‘idealistic, networked professionals of reflexive modernity’ (p. 4) can more easily exert normative control over these organizations. Although Whittington and Yakis-Douglas describe this process as one of stakeholder pressure and public shaming, their causal theorizing comes with an optimistic outlook: although the ‘corporate subpolitics’ (p. 5) of transparency and accountability can be strenuous at times, they eventually lead to organizations that are economically sustainable and are able to meet their stakeholders’ demands for ethically appropriate business practices. Typical for the causal perspective, the authors argue that while transparency is necessary for achieving these two goals, it needs to be calibrated to unleash its full potential, i.e. by combining it with well-dosed forms of inclusion into strategy-making, too.

Bringing together an intuitive metaphor with a business case and ‘actionable’ insights, the causal perspective has become widely popular among practitioners and scholars alike. However, the success of the causal perspective has also attracted critics. As shown in the introductory case of Google, not all stakeholders are convinced that greater transparency on AI technology will increase Google’s accountability regarding the (unexpected) outcomes of this powerful technology. Quite to the contrary, it seems, Google’s efforts to increase transparency on AI did not lead to a generative dialogue with its stakeholders, but the creation of PAI has fuelled even more wide-ranging and persistent protest. In addition to this practical example, there is growing concern among researchers regarding the optimism that transparency leads to accountability. In a recent review of the literature on the relationship between transparency and accountability in the public sector, Cucciniello and colleagues (2017, p. 40) find that only a minority of studies can firmly establish a causal relationship and that several studies found ‘no relationship between transparency and accountability at all’. Taken together, these practical and theoretical inconsistencies have given rise to an alternative perspective on transparency and accountability.

Critical Perspective: Transparency Clouds Accountability

Like the causal perspective, a critical perspective on transparency and accountability assumes a meaningful relationship between the two concepts. However, research from a critical perspective believes this relationship to be complex rather than direct. Due to the many facets of visibility in and around organizations, transparency can never lead to universal but only to limited accountability – no matter how much it is calibrated or tweaked. While the root metaphor of the causal perspective is a window through which sunlight wipes out ‘bacterial’ wrongdoing in organizations, the critical perspective understands transparency as a prism which refracts organizational accounts in manifold and complex ways (Flyverbom, 2020). Critical research on transparency and accountability is conducted from many different theoretical and methodological vantage points. It is this plurality of approaches which contributes to the growing popularity of this perspective, as it confirms its key message: the relation between transparency and accountability is complex and cannot adequately be described through a single theory or method. Briefly introducing two lines of work within the critical camp shows how a diversity of theories and methods underpins the unity of the perspective.

One line of critical work argues that transparency can only lead to incomplete accountability. For example, one way in which such incompletion comes about is when organizations ‘decouple’ (Meyer & Rowan, 1977) their secretive core operations from their outward-facing facade of transparency. In his study on data-centred transparency in local government agencies, Heimstädt (2017) found that city employees facing stakeholder pressure to release their internal data sets to the public, carefully manipulated, curated and remixed data sets in order to protect the underlying processes and services (see also Denis & Goëta, 2017; Kornberger, Meyer, Brandtner, & Höllerer, 2017). Another way in which critical studies explain incomplete accountability is through the mediated nature of transparency. These studies understand transparency as a form of visibility that is inevitably mediated by a technology or tool. While there are different tools at hand, each of them can only make visible some aspects of an organization while leaving others out (e.g. Ananny & Crawford, 2018; Hansen, 2015; Hansen & Weiskopf, 2019). For example, ‘disclosure devices’ (Hansen & Flyverbom, 2015) such as documents and personal meetings used in a due diligence process produce transparency and accountability in a narrative form that is closely bound to individuals and the specific context of its production. On the other hand, quantitative devices such as numbers, ratings and rankings produce a form of transparency and accountability that is more mobile than the narrative accounts of due diligence.

Another line of critical work suggests that transparency can also be detrimental to the ideal of accountability. Much of the research revisits Marilyn Strathern’s (2000, p. 310) famous question ‘what does visibility conceal?’ in the context of digitally networked environments. In his study on the nascent Pirate Party in Germany, Ringel (2019) describes how party members tried to realize radical transparency to reach their goal of universal accountability. In order to live up to this ideal of radical transparency, party members opened their mailing lists for outsiders, uploaded all documents produced in their working groups to the Internet, and video streamed their caucus meetings. Just weeks into this phase of radical transparency, tensions surfaced, which quickly pushed the organization to the brink of collapse. The crisis was eventually overcome when members reintroduced formal and informal ‘boundaries of visibility’ (Ringel, 2019, p. 709). The case of the Pirate Party demonstrates that instead of creating greater accountability for an organization’s activities, radical transparency can paralyse an organization, thereby hindering the production of any account-worthy activities in the first place.

The critical perspective on transparency and accountability has developed as a counterpoint to the causal perspective. Sometimes more or less directly, critical studies accuse the causal perspective of uncritically adopting and perpetuating transparency as a political ideology instead of providing an ‘honest’ assessment of what it is and how it relates to accountability (Zyglidopoulos & Fleming, 2011). Proponents of the causal perspective have tried to deflect this criticism by calibrating and complementing transparency in their models, e.g. by also highlighting the important role of inclusion for accountable strategy-making (Whittington & Yakis-Douglas, 2020). In addition, proponents of the causal perspective have argued that critical research has little practical use and is only interested in ‘debunking’ proposals on how to curb excessive corporate power – without offering any alternative. Without doubt there has been a generative momentum in this back and forth between causal and critical perspectives, helping both of them to refine and expand their arguments. However, when taking a step back and reflecting about the commonalities rather than the differences between these well-practised antagonists, new limitations of their explanatory potential come into sight.

Towards a Constructive Perspective: Turning Resources into Topics

Causal and critical perspectives have in common that they give priority to definitions of transparency and accountability, which are developed by researcher-analysts and mostly disregard the ways in which people in and around organizations (i.e. practitioner-analysts) draw on and come to terms with these concepts. Causal and critical studies either synthesize existing scientific definitions or justify why they favour one scientific definition over another. For example, when a causal study operationalizes the transparency–accountability link in the same way as other studies, they can more easily demonstrate the robustness of their findings (Cucciniello et al., 2017; Schnackenberg & Tomlinson, 2016). Although they do so more implicitly than causal studies, most critical studies draw on researcher-analyst categories of transparency and accountability, too. For example, when researchers describe an organization’s transparency efforts as a form of ‘decoupling’ (Heimstädt, 2017), ‘translation’ (Kornberger et al., 2017) or ‘obfuscation’ (Fabrizio & Kim, 2019), they picture a researcher-analyst ideal of ‘coupled’, ‘immediate’ or ‘illuminated’ transparency from which their empirical observations deviate.

Pointing out this commonality is not just an intellectual sleight of hand but surfaces the limitations of the existing perspectives to fully understand cases such as Google’s use of AI technology. From a causal perspective, Google’s successful strategic shift towards AI since 2017 seems to confirm the hypothesis that more transparency leads to greater accountability and legitimacy. From a critical perspective, the protests against Google’s unethical use of AI can be interpreted as proof that their work with PAI is not more than ‘openwashing’ (Heimstädt, 2017) and that the corporation failed to create ‘real’ transparency. While we have no doubt that these interpretations are of interest to public and scholarly debate, we believe that this interest comes at a price. In order to make clear-cut statements whether transparency and accountability ‘work’, these perspectives deliberately exclude from their analysis the question of what actually happens when people try to come to terms with the concepts of transparency and accountability in their everyday work. To illuminate this blind spot we propose a third, constructive perspective.

A constructive perspective on transparency and accountability turns these analytic resources for studying empirical phenomena into topics of inquiry and hence phenomena in their own right. Through this move, the constructive perspective becomes able to sidestep what ethnomethodologist Melvin Pollner (2012, p. 14) described as the ‘topic–resource confusion’, in which ‘the analyst “naively” employs the very assumptions, practices, categories and concerns of host participants’ and eventually ‘becomes conceptually ensnared in assumptions and practices which might otherwise be treated as phenomena’ (see also Zimmerman & Pollner, 1970). Causal and critical perspectives understand the popularity of transparency and accountability in organizational life as a mandate to use these concepts as analytic resources to explain organizational life. In contrast, research from a constructive perspective studies how people in organizations use transparency and accountability as analytic resources in their everyday work and with what consequences. Rather than validating or rebuffing the effect of transparency on accountability, such a perspective allows researchers to understand how the construction of these categories is entwined with relations of power in everyday activities. Rather than making any assumptions about a pre-existing relation between transparency and accountability (direct or complex), a constructive perspective allows researchers to observe how analysts create and configure relations between transparency and accountability.

Our proposal for a constructive perspective is informed by emerging, yet previously disconnected lines of work on transparency and accountability. One emerging line of work examines transparency as performative practice. For example, drawing on the concept of ‘non-performativity’ (Ahmed, 2012), Dobusch and Dobusch argue that if an organization’s explicit commitment to transparency gets treated as if it was doing transparency, this commitment can lead towards not doing transparency at all (Dobusch & Dobusch, 2019). Although strictly speaking the idea of ‘not doing transparency at all’ treats transparency as a researcher-analyst concept as well, their argument highlights that for understanding how an organization manages its visibilities it can be very important to examine how people in that organization lay claim to and contest the concept of transparency.

Even more instructive for a constructive perspective are studies that build on the ethnomethodological impetus to ‘transform everyday categories and objects into activities that constitute them’ (Samra-Fredericks, 2015, p. 478) and have used this lens to examine the notion of accountability. In one study from this line of work, Neyland (2016, p. 53) reports from a project in which several large organizations collaboratively developed a video-based surveillance system for public spaces, and to which he was ‘invited to oversee the ethics of the technology under development’. Turning the analytic resource of accountability into a topic of inquiry, he observed and reported how the ‘account-able order’ (p. 55) for different features of the system was established and subject to change over time. For example, he describes closely how project members explained to him and discussed among each other if and how the surveillance system produces accounts of ‘abandoned luggage’. In his role as an embedded ethicist, Neyland repeatedly presented this and other observations to the project’s ethics board. The board, he describes, would then draw on his accounts when presenting itself as accountable towards the press and the public. Table 1 summarizes similarities and differences of the three perspectives on transparency and accountability.

Table 1.

Three perspectives on transparency and accountability.

Perspective	Causal	Critical	Constructive
Explanatory potential	How transparency can create accountability	Why transparency can only create limited accountability	What is at stake when people claim and contest transparency and accountability
Status of concepts	Concepts are analytic resources used by researcher-analysts	Concepts are analytic resources used by researcher-analysts	How practitioner-analysts use concepts as analytic resources becomes the topic of inquiry for researcher-analysts
Relation between concepts	Direct relation	Complex relation	Relating as something practitioner-analysts do and researcher-analysts observe
Exemplary study	Whittington & Yakis-Douglas (2020)	Ringel (2019)	Neyland (2016)

Sites of Ethical Contestation: Studying Transparency and Accountability in Action

The constructive perspective requires new and creative methodological strategies that allow researchers to study how people engage accountability and transparency in their everyday activities, that is, ‘the ways in which members assemble particular scenes so as to provide for one another evidences of a social order [e.g. that of transparency or accountability] as-ordinarily-conceived’ (Zimmerman & Pollner, 1970, p. 83). In this final section, we therefore introduce sites of ethical contestation as a methodological strategy that helps organization scholars to study transparency and accountability beyond causal and critical perspectives as mundane accomplishments and means of organizing in themselves.

We define sites of ethical contestation as spaces in which new and ambiguous matters of transparency and accountability are claimed, contested and configured. While ethical concerns around transparency and accountability might surface anywhere in an organization, these spaces are characterized by their lack of ‘ontological security’ (Ziewitz, 2019, p. 713) regarding these concepts. For example, we can assume that people in a CSR department regularly have heated debates about issues like transparency indicators or accountability standards. However, we argue that CSR departments are not sites of ethical contestation as people there generally agree on the ontological nature of transparency and how it is linked to accountability. In what we define as a site of ethical contestation, people struggle much more with the question of what transparency and accountability actually are (for example, in relation to an emergent phenomenon such as AI). This lack of ontological security requires people to explicate, define and defend their definitions vis-a-vis other participants and hence make these controversies available for analysis. Looking at these sites makes visible how transparency and accountability are not only problems for organization scholars and other scientists who study them, but also for those people who engage in their construction in the first place.

We suggest that a promising place to look for these sites is at the fringes of organizations. While organizations are fairly good at creating ontological stability within their boundaries, it is at the interfaces with their environments where ontological struggles unfold. A more traditional example for a site of ethical contestation is a demonstration. With the rise of the ‘gig economy’, platform organizations claim that their practices of algorithmic management create an unprecedented level of transparency in the labour market. To contest this claim, gig workers more recently began heading to the streets and demonstrating against what they perceive to be an exploitative opacity of these systems (O’Connor, 2016). For organization theorists, demonstrations, strikes or even shareholder meetings are rather well-known sites for studying conflict. However, the new methodological concept also points towards less obvious settings that are even more specific to controversies around transparency and accountability. For example, situations in which whistleblowers interact with investigative journalists and debate the nature of their (intended) disclosure can be studied as sites of ethical contestation. In her documentary film Citizenfour Laura Poitras provides a glimpse into the eight-day period during which Edward Snowden and a handful of journalists met in a hotel room in Hong Kong to prepare the first media reports on the NSA’s mass surveillance activities. Studying this setting as site of ethical contestation can help us to better understand how the involved participants navigated this situation of unprecedented uncertainty. In the following we introduce three tactics that can help organization scholars to study these sites and thereby harness the potential of the constructive perspective.

Tracing provisional practices

The lack of ontological security at sites of ethical contestation invites testing and experimentation with alternative versions of transparency and accountability. One tactic to study these concepts from a constructive perspective is hence to trace provisional practices over time. The idea of these sites as spaces in which uncertainty fosters generativity is informed by what Furnari (2014) has called ‘interstitial spaces’. When actors from adjacent fields come together in informal settings such as hobbyist clubs, hangouts or meet-ups, he argues, their interactions might result in innovative social practices. Tracing provisional practices allows organization scholars to better understand how different versions of transparency and accountability are piloted, tested, refined or discarded along the way. An example for sites with a culture of collective experimentation around transparency and accountability are ‘civic hacknights’ (Heimstädt & Dobusch, 2018; Heimstädt & Reischauer, 2019). At civic hacknights, civil servants and engaged citizens from different professions and domains meet after work for a few hours to collaboratively work on social problems. Oftentimes these problems involve matters of transparency and accountability. A typical type of solution to such problems are software prototypes such as apps that are supposed to render local government more transparent by visualizing government datasets (Schrock, 2016). Tracing how such practices and prototypes change over time allows organization scholars not to answer the question of whether transparency leads to accountability or not, but how people negotiate underlying social problems through the concepts of transparency and accountability.

Comparing staged performances

Another tactic to examine sites of ethical contestation is to compare staged performances such as keynotes or panel discussions. Research on field-configuring events has demonstrated how such ‘ceremonial and dramaturgical activities’ (Lampel & Meyer, 2008, p. 1027) can help organizations to cope with collective uncertainty and to provide direction for the development of an entire field (Möllering & Müller-Seitz, 2018). A site at which staged performances play an important role for Google’s AI strategy is the Annual Workshop on Fairness, Accountability and Transparency in Socio-Technical Systems (FAT*). Although the FAT* workshops comes across as regular academic conferences at first sight (no-frills website, paper presentations, keynotes), it brings together a broad array of actors including activists, journalists and industry representatives. It is this heterogeneity of actors that makes FAT* different from the relative ontological security of a traditional academic conference. At the 2020 FAT* workshop in Barcelona, the Google-related PAI became involved as well and staged three different performances. First, members of the non-profit presented a ‘conference paper’ on the explainability of machine learning. Second, PAI hosted an ‘implications tutorial’ in which the matters of transparency and accountability were addressed through design-thinking methods. Third, the organization invited participants to an ‘interactive happy hour’ at which they could ‘mingle’ with PAI staff (Partnership on AI, January 21, 2020). Comparing the way transparency and accountability have been mobilized across these performances can illuminate how different audiences grapple with the uncertainty involved in emerging technologies such as AI.

Researching the researchers

Finally, a focus on sites of ethical contestation invites organization scholars to develop greater reflexivity about the role of scientific expertise in the enactment of transparency and accountability. In the causal and critical perspective, scientific experts figure as ‘neutral’ researcher-analysts, who observe the interplay between organizations and their stakeholders from a distance. The new methodological strategy allows organization scholars to examine how and to what effect scientists are enrolled in organizational struggles around transparency and accountability. The sociologist Gil Eyal has recently described this amalgamation of science and society through the concept of ‘trans-scientific spaces’ (Eyal, 2019). Trans-scientific spaces, he argues, are temporary situations in which pressing concerns are discussed and negotiated, but which are neither located in the realm of ‘society’ nor in the realm of ‘science’ alone. In trans-scientific spaces, scientists are confronted with questions which ‘should and will be asked of science’ – as they are questions of facts – ‘even though scientists cannot answer them’ as these questions transcend the practical abilities and apparatuses of scientists (Eyal, 2019, p. 257).

We are returning one last time to our introductory vignette about Google’s AI strategy. Interpreting sites of ethical contestation as trans-scientific spaces directs our attention towards the ways in which Google tries to create ontological security by turning the questions of transparency and accountability into a question of scientific judgement. In March 2019, Google created an external advisory board to monitor its use of AI, yet shut it down only a few days later in response to severe public criticism about its composition (alongside scientists, the firm also included a conservative policy-maker). As the short-lived board began to fall apart, one of the scientific experts justified his decision to step down on Twitter: ‘While I’m devoted to research grappling with key ethical issues of fairness, rights, and inclusion in AI, I don’t believe this is the right forum for me to engage in this important work’ (Statt, 2019). From this episode and brief statement, many new questions for organization scholars arise. How do organizations try to enrol scientists in order to address issues of transparency and accountability? How do these scientists navigate the tenuous middle ground between societal and organizational interests on the one hand and professional ethics on the other? When does a search for ethical business practices turn into a matter of ethical concern in itself?

Conclusion

Transparency and accountability have become major concerns for organization scholars. Some of them are sceptical that greater transparency of firms means that these firms can be held more accountable for their activities. Others argue that firms will conform to public expectations when exposed to greater scrutiny through this public. Even more optimistically, some scholars argue that when powerful technology firms such as Google engage in more transparent forms of strategy-making, this openness shifts some of their excessive power (e.g. regarding issues like data, privacy and infrastructure) back to consumers and other stakeholders (Whittington & Yakis-Douglas, 2020). In this essay, we have argued that these perspectives too closely replicate the lines of argumentation developed by practitioners in and around such powerful organizations. While this proximity makes causal and critical studies easily digestible for practitioners, their ability to produce novel and surprising insights for organization scholars is somewhat limited.

We therefore propose the constructive perspective as an alternative vista on transparency and accountability. A constructive perspective deliberately blurs the distinction between researchers and practitioners. Instead of describing what people in organizations do through the analytic concepts of transparency and accountability, organization scholars who adopt a constructive perspective study how people in organizations draw on transparency and accountability when analysing and organizing their everyday work. A constructive perspective hence turns the analytic resources of the causal and critical perspective into topics of inquiry in their own right. Exploring sites of ethical contestation as a new methodological strategy, organization scholars will not only be able to produce surprising theoretical insights for other researchers, but will also help practitioners observe their engagements with transparency and accountability in novel ways.

Footnotes

Acknowledgements

We thank David Seidl and Joep Cornelissen for their helpful comments on this paper. We also thank Richard Whittington and Basak Yakis-Douglas for being such generative sparring partners in this controversy. Finally, we are grateful to Malte Ziewitz for his insightful feedback on an earlier version of this paper.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author biographies

Maximilian Heimstädt is head of a research group at the Berlin-based Weizenbaum Institute and an affiliated senior researcher (‘Habilitand’) at Witten/Herdecke University. Drawing on analytic resources from organization theory and STS theory, he studies new forms of organizing in digitally networked environments.

Leonhard Dobusch is Professor of Organization at the Department of Organization and Learning, Universität Innsbruck, Austria and Scientific Director of the Vienna-based Momentum Institute. His research interests include the management of digital communities, private regulation via standards and openness as an organizing principle.

References

Ahmed

(2012). On being included: Racism and diversity in institutional life. Durham, NC: Duke University Press.

Albu

O. B.

Flyverbom

(2019). Organizational transparency: Conceptualizations, conditions, and consequences. Business & Society, 58, 268–297.

Ananny

Crawford

(2018). Seeing without knowing: Limitations of the transparency ideal and its application to algorithmic accountability. New Media & Society, 20, 973–989.

BBC News. (2018, April 5). Google staff challenge Pentagon project. BBC News. Retrieved from https://www.bbc.com/news/business-43656378

Bernstein

(2017). Making transparency transparent: The evolution of observation in management theory. Academy of Management Annals, 11, 217–266.

Brandeis

L. D.

(1914). Other people’s money and how the bankers use it. New York: Frederick A. Stokes.

Cucciniello

Porumbescu

G. A.

Grimmelikhuijsen

(2017). Twenty-five years of transparency research: Evidence and future directions. Public Administration Review, 77, 32–44.

Denis

Goëta

(2017). Rawification and the careful generation of open government data. Social Studies of Science, 47, 604–629.

Dobusch

(2019). The relation between openness and closure in open strategy: Programmatic and constitutive approaches to openness. In Seidl

von Krogh

Whittington

(Eds.), The Cambridge handbook of open strategy (pp. 326–336). Cambridge: Cambridge University Press.

10.

Dobusch

Müller-Seitz

(2019). Closing for the benefit of openness? The case of Wikimedia’s open strategy process. Organization Studies, 40, 343–370.

11.

Eyal

(2019). Trans-science as a vocation. Journal of Classical Sociology, 19, 254–274.

12.

Fabrizio

K. R.

Kim

E.-H.

(2019). Reluctant disclosure and transparency: Evidence from environmental disclosures. Organization Science, 30, 1207–1231.

13.

Flyverbom

(2020). The digital prism: Transparency and managed visibilities in a datafied world. New York: Cambridge University Press.

14.

Furnari

(2014). Interstitial spaces: Microinteraction settings and the genesis of new practices between institutional fields. Academy of Management Review, 39, 439–462.

15.

Gegenhuber

Dobusch

(2017). Making an impression through openness: How open strategy-making practices change in the evolution of new ventures. Long Range Planning, 50, 337–354.

16.

Haack

Martignoni

Schoeneborn

(2020). A bait-and-switch model of corporate social responsibility. Academy of Management Review.

17.

Hansen

H. K.

(2015). Numerical operations, transparency illusions and the datafication of governance. European Journal of Social Theory, 18, 203–220.

18.

Hansen

H. K.

Flyverbom

(2015). The politics of transparency and the calibration of knowledge in the digital age. Organization, 22, 872–889.

19.

Hansen

H. K.

Weiskopf

(2019). From universalizing transparency to the interplay of transparency matrices: Critical insights from the emerging social credit system in China. Organization Studies. doi https://doi.org/10.1177/0170840619878474

20.

Heimstädt

(2017). Openwashing: A decoupling perspective on organizational transparency. Technological Forecasting and Social Change, 125, 77–86.

21.

Heimstädt

Dobusch

(2018). Politics of disclosure: Organizational transparency as multiactor negotiation. Public Administration Review, 78, 727–738.

22.

Heimstädt

Reischauer

(2019). Framing innovation practices in interstitial issue fields: Open innovation in the NYC administration. Innovation: Organization & Management, 21, 128–150.

23.

Kornberger

Meyer

R. E.

Brandtner

Höllerer

M. A.

(2017). When bureaucracy meets the crowd: Studying ‘open government’ in the Vienna city administration. Organization Studies, 38, 179–200.

24.

Lampel

Meyer

A. D.

(2008). Field-configuring events as structuring mechanisms: How conferences, ceremonies, and trade shows constitute new technologies, industries, and markets. Journal of Management Studies, 45, 1025–1035.

25.

Meyer

J. W.

Rowan

(1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340–363.

26.

Möllering

Müller-Seitz

(2018). Direction, not destination: Institutional work practices in the face of field-level uncertainty. European Management Journal, 36, 28–37.

27.

Neyland

(2016). Bearing account-able witness to the ethical algorithmic system. Science, Technology, & Human Values, 41, 50–76.

28.

O’Connor

(2016, September 8). When your boss is an algorithm. Financial Times. Retrieved from https://www.ft.com/content/88fdc58e-754f-11e6-b60a-de4532d5ea35

29.

Partnership on AI. (2020, January 21). Join the Partnership at FAT* 2020. Retrieved February 25, 2020, from The Partnership on AI website: https://www.partnershiponai.org/pai-at-fat-star/

30.

Pollner

(2012). The end(s) of ethnomethodology. American Sociologist, 43, 7–20.

31.

Ringel

(2019). Unpacking the transparency–secrecy nexus: Frontstage and backstage behaviour in a political party. Organization Studies, 40, 705–723.

32.

Samra-Fredericks

(2015). Researching everyday practice: The ethnomethodological contribution. In Golsorkhi

Rouleau

Seidl

Vaara

(Eds.), Cambridge handbook of strategy as practice (pp. 477–490). Cambridge: Cambridge University Press.

33.

Schnackenberg

A. K.

Tomlinson

E. C.

(2016). Organizational transparency: A new perspective on managing trust in organization–stakeholder relationships. Journal of Management, 42, 1784–1810.

34.

Schrock

A. R.

(2016). Civic hacking as data activism and advocacy: A history from publicity to open government data. New Media & Society, 18, 581–599.

35.

Statt

(2019, April 4). Google dissolves AI ethics board just one week after forming it. Retrieved February 24, 2020, from The Verge website: https://www.theverge.com/2019/4/4/18296113/google-ai-ethics-board-ends-controversy-kay-coles-james-heritage-foundation

36.

Strathern

(2000). The tyranny of transparency. British Educational Research Journal, 26, 309–321.

37.

Whittington

(2019). Opening strategy: Professional strategists and practice change, 1960 to today. Oxford and New York: Oxford University Press.

38.

Whittington

Cailluet

Yakis-Douglas

(2011). Opening strategy: Evolution of a precarious profession. British Journal of Management, 22, 531–544.

39.

Whittington

Yakis-Douglas

(2020). The grand challenge of corporate control: Opening strategy to the normative pressures of networked professionals. Organization Theory, 1. https://doi.org/10.1177/2631787720969697

40.

Whittington

Yakis-Douglas

Ahn

(2016). Cheap talk? Strategy presentations as a form of chief executive officer impression management. Strategic Management Journal, 37, 2413–2424.

41.

Wong

J. C.

(2019, October 3). Google reportedly targeted people with ‘dark skin’ to improve facial recognition. The Guardian. Retrieved from https://www.theguardian.com/technology/2019/oct/03/google-data-harvesting-facial-recognition-people-of-color

42.

Yakis-Douglas

Angwin

Ahn

Meadows

(2017). Opening M&A strategy to investors: Predictors and outcomes of transparency during organisational transition. Long Range Planning, 50, 411–422.

43.

Zerega

(2018, May 11). Google’s use of AI to mimic humans is unethical and bad UX. Retrieved February 18, 2020, from VentureBeat website: https://venturebeat.com/2018/05/11/googles-use-of-ai-to-mimic-humans-is-unethical-and-bad-ux/

44.

Ziewitz

(2019). Rethinking gaming: The ethical work of optimization in web search engines. Social Studies of Science, 49, 707–731.

45.

Zimmerman

D. H.

Pollner

(1970). The everyday world as a phenomenon. In Douglas

(Ed.), Understanding everyday life (pp. 80–104). Chicago: Aldine.

46.

Zyglidopoulos

Fleming

(2011). Corporate accountability and the politics of visibility in ‘late modernity’. Organization, 18, 691–706.

Transparency and Accountability: Causal,Critical and Constructive Perspectives

Abstract

Keywords

Introduction

Causal Perspective: Transparency Creates Accountability

Critical Perspective: Transparency Clouds Accountability

Towards a Constructive Perspective: Turning Resources into Topics

Sites of Ethical Contestation: Studying Transparency and Accountability in Action

Tracing provisional practices

Comparing staged performances

Researching the researchers

Conclusion

Footnotes

Acknowledgements

Funding

Author biographies

References