Introduction: Into the goldmine

October 2017, Brazil: A report by journalist Bruno Abbud of the cultural magazine Piauí exposed private conversations in a WhatsApp chat of a right-wing group called Movimento Brasil Livre (see Messenberg, 2017). Among other things, the group discussed the behind-the-scenes meetings of their members with national politicians and leaders, now accessible to public purview in a total of over 685 pages of transcripts dealing with health, public policy, and education, among others (Abbud, 2017).¹ This is just one of the many instances in which trust between participants in chat apps of a political nature is misused, no matter what the reasons. More and more scholars are devoting their attention to political activism facilitated by real-time chat applications like Whatsapp and Telegram (see, among others, Barbosa, 2017; Alimardani and Milan, 2018; Potnis, Demissie, Trimmer and Cleek, 2018; Agur and Frisch, 2019). We can safely argue that messaging apps are today one of the primary sites for media activism to emerge and unfold (see, e.g. Pickard and Yang, 2017). However, and despite the fact they represent a real goldmine for scholars, messaging apps have long remained under the radar of scholars of social movements and digital activism, largely because of the limited permeability of chat groups to overt and covert observation alike. The many ‘closed’ groups mobilizing on chat apps are typically difficult to penetrate; their existence is often unknown to scholars. In addition, the high frequency of exchanges makes digital ethnography a 24/7 exercise. Nonetheless, the popularity of chat apps for political debate and the coordination of collective action makes them a fruitful venue for research – and more and more scholars are taking advantage of this amazing opportunity. But, as the opening example shows, things might be more complicated than they appear at first sight. The recent introduction of end-to-end encryption to chat apps, then, risks giving activists a false sense of security that encourages individuals and groups to use these channels also for sensitive exchanges, thereby exposing participants to potential risks that researchers might indirectly contribute to amplifying. It is precisely this centrality of chat apps in contemporary activism and the perceived secrecy surrounding the space that studying interactions in what we can imagine as a ‘(private) public sphere’ (Myers West, 2017) requires carefully-negotiated ethical commitments.

This article emerges out of a common interest of the authors, sparked by a conversation that followed the presentation of Barbosa’s study of #UnidosContraOGolpe (UCG, United Against the Coup). This mobilisation, which emerged in Brazil in the aftermath of the impeachment of then-President Dilma Rousseff (2016), was facilitated by a private group on WhatsApp. As we will explain, Barbosa embarked on a digital ethnography of group interaction. Milan, a long-term student of the relation between movements and technology (2013), was intrigued by this new genre of media activism, which expands the repertoire of digital activism and encapsulates the political potential of chats apps as a civic engagement tool. In particular, she wondered about the potential ethical ramifications of doing research on chat apps, and private groups in general – and the two researchers started to exchange ideas. What follows is the result of this ongoing exchange and intends to stimulate a much-needed collective reflection on the subject matter. What are the limits of (c)overt participant observation in this emerging ‘(private) public sphere’? How can we safeguard the ethics of research, protecting user privacy and respecting the intimacy of their exchanges, while simultaneously taking advantage of such a rich data source? As we shall see, the nature of the medium in question, and of the interactions between group participants lend themselves particularly well to covert observation, in ways that ‘traditional’ observation in ‘real’ life typically does not. Even when one discloses their researcher identity, the fluidity of belonging and participation in a chat app creates a situation that it is not really possible to inform every group participant about the ongoing observation, let alone remind each of it; ensuring ‘informed’ consent is probably just wishful thinking in very large groups. Deceiving participants is thus a concrete risk and a tempting possibility. What’s more, the informality of the space and the confidentiality of conversations, mediated by an intimate technology like the smartphone, exposes ordinary users to more risks that they might themselves imagine. It is a slippery slope that researchers can no longer ignore.

The article investigates the methodological and ethical challenges of conducting research of an ethnographic nature on WhatsApp. Considering chat apps as simultaneously a social phenomenon and a tool with specific affordances for political participation as well as for research, we reflect on the ethics of qualitative data collection when the field site is an encrypted chat app such as WhatsApp, Telegram and Signal. In particular, we focus on the possibilities of digital ethnography in private groups and offer a viable approach to thinking through the ethics of doing research in this digital space. The article builds on and brings into dialogue findings from two research projects. As mentioned above, Barbosa has investigated interactions on the #UnidosContraOGolpe private group on Whatsapp (2017), while Milan has been collecting data on the evolution of political participation and social media usage amongst grassroots activists in various countries since 2012, as part of an ongoing project on ‘cloud protesting’ (see, inter alia, S. Milan, 2015, 2018 and on ‘data activism’ Milan, 2017).

The paper is composed of four parts. Firstly, we present what is at stake for researchers in encrypted chat apps, focusing on the diffusion of WhatsApp and its siblings worldwide and their role in digital and media activism today, and showing, data in hand, why this represents a great research opportunity. Secondly, we explore the case study that triggered the conversation and inspired our joint interrogation into the ethics of research on chat apps. Thirdly, we reflect on digital ethnography as an adaptive method to study interactions in chat apps, reflecting in particular on the risks that come with it. We then offer our considerations for thinking and practising ethics in chat app research, building also on existing literature on research with politically active subjects.

The blast of WhatsApp and its siblings

Why should we care about Whatsapp and other similar platforms in the study of media activism? WhatsApp is a cross-platform messaging app allowing users to exchange messages over a phone’s data traffic without paying extra for short-text messaging. Launched in 2009, it turns ten year old in 2019 and is currently the world’s fastest growing platform. It became a global phenomenon in 2012; Facebook bought it for USD 19 billion in 2014. Over the years, its functionalities were updated several times. Today a WhatsApp user can share pictures, send files, and call other users, including video and/or voice and group calls. She can also engage in the very popular ‘private group chats’ that are analysed in this article.

By the end of 2017, WhatsApp counted 1.5 billion active monthly users (Garimella and Tyson, 2018). It is the preferred app in more than 100 countries around the world (Constine, 2018), with high penetration rates in countries as diverse as India, Indonesia, Malaysia, Mexico, Brazil and South Africa (Sevitt, 2017). According to a research by the company We Are Social, Whatsapp is the top social messenger app in Central and Latin America, Canada, most of Africa and Europe, and Russia (Iqbal, 2019). Whatsapp is pursued by Facebook Messenger with 1.3 million active users, the Chinese multi-purpose app WeChat and Tencent QQ (respectively 1 million and 803,000 users), followed by Skype, Snapchat, Viber, LINE and Telegram, with a mere 200,000 users (Statista, 2019) – in other words, the large majority of smartphone holders have installed one or more of these applications.

In recent years, several of these services have adopted end-to-end encryption, in response to public concerns about state snooping as uncovered by the Snowden revelations (for an overview, see Greenwald, 2014). End-to-end encryption (the end indicates the ‘endpoint’, meaning the client device) prevents the hosting server or any intercepting third party from accessing the content of the message (Ermoshina, Musiani, and Halpin, 2016).² Following the example of, among others, Signal – an open-source, encrypted messaging service spearheaded in 2014 by Open Whisper Systems (now the non-profit Signal Foundation) – WhatsApp has implemented end-to-end encryption in April 2016 – although not without trouble, as reported by Santos and Faure (2018). Technically, it means that messages remain private. Yet, WhatsApp’s end-to-end encryption does not necessarily imply that Facebook is not in a position to collect WhatsApp data (Zanon, 2018), in what has been called ‘a metadata imbroglio’ (Brewster, 2017), whereby some data – such as a user’s phone number, contact list, and usage data – are shared with the ‘mother’ company Facebook.³ Table 1 summarises the main features of the most popular messaging apps.

Chat apps play a key role in contemporary activism, including media activism (see, e.g. Custódio, 2017; F. L. F. Lee and Chan, 2016; Caetano et al., 2018; Treré, 2018; Akbari and Gabdulhakov, 2019). Despite it being frequently considered a flywheel of fake news, especially in countries of the so-called Global South (Burgos, 2019; cf. Marda and Milan, 2018), WhatsApp has established itself as a powerful political tool for spreading political information (Montag et al., 2015). In Brazil, for example, it has been central in the organisation of the 2018 truck drivers’ strike, considered the largest in the history of the country (Fox, 2018). But it has also acted as a ‘weaponized’ chat app for the dissemination of misinformation during the 2018 Brazilian Presidential elections (Pereira and Bojczuk, 2018; see also Nemer, 2018; Córdova and Barbosa, 2019). Despite these potentially contradictory manifestations, it is increasingly popular among ordinary users in Brazil (Spyer, 2017) as well as politically-active individuals. Little research however has focused on Whatsapp as such.⁴ Next, we provide an example of how ethical considerations played out in studying interactions on Whatsapp in the absence of codified ethical guidelines.

#UnidosContraOGolpe: The challenge of studying social interactions in a private chat

Brazil is one of the forerunners as far as WhatsApp penetration is concerned, with 120 million active users (Digital Information World, 2019). As reported by the WhatsApp Nation study (2016), 76 per cent of messaging app users regularly use ‘Zap Zap’, as Brazilians call Whatsapp (Vieira, 2018). According to the 2017 Digital News Report, 46 per cent of Brazilian citizens use WhatsApp to find news (Reuters Institute, 2017). With such a prominent role in a country that is rightly considered one of the cradles of media activism in Latin America, Barbosa set out to observe how the app enabled activism during a difficult moment in the recent history of the country: the controversial impeachment of its president.

#UnidosContraOGolpe (UCG) was a group of leftist activists and concerned citizens created on WhatsApp to oppose the impeachment of Rousseff. In fact, the impeachment was perceived by many Brazilians as an unwarranted soft ‘coup’ (Souza, 2016; Weber, Bastone and Barbosa, 2018). Participants advocated the return of Rousseff to the Presidency of the Republic. The individuals enrolled in UCG, a ‘private’ group with ‘by invitation only’ membership, expressed indignation with the political situation and discussed potential actions and alternative scenarios. The confrontation between diverse identities in the group allowed civic engagement to flourish in this space. Most users were also active protesting in the streets, but within WhatsApp they created a new style of political participation able to connect the square and the digital sphere. The diverse opinions among users grew alongside with interaction on the chat, resulting in a novel, shared collective voice – as observed in earlier forms of media activism (Hackett and Carroll, 2006; Meikle, 2002; S. Milan, 2013, 2017).

How did UCG come about? The group was created on 30 March 2016 by a concerned citizen from Florianópolis, in the South of the country, around the time in which the Parliament Lower Chamber was called to vote on the request of impeachment against Rousseff.⁵ Barbosa became a member of the group and performed an extensive digital ethnography, logging information daily in order to investigate member interaction. At the time of joining, the group counted 256 registered users (the maximum allowed by the service), with the exact number varying from day to day; interestingly, the vast majority of members were listed also as ‘administrators’, which allowed them to add new members or exclude existing ones, signalling the absence of identified leaders.⁶ However, guidance with respect to the ethical implications of a research approach of this kind was not readily available locally to our researcher. To begin with, the hosting institution did not require an ethical clearance – as it is often the case in countries outside of the Anglo-Saxon context. Available ethical guides and indications (e.g., Association of Internet Researchers, 2002, 2016; Chesters, 2012; S. Milan, 2014) did not yet provide ready-to-use insights on the specific app, as obviously they do not have the capacity to keep up with the pace of all tech innovations and user adoption. In the lack of codified codes of conducts, Barbosa embarked in a process of moral decision-making.

The first challenge was of methodological nature: how to develop a creative approach to digital ethnography that did not harm or interfere with the interactions among chat members? Upon asking to join the private group in April 2016 and being immediately admitted by the administrators (those controlling the ‘gates’ of the channel), Barbosa posted a message to inform group members of the research and to announce that some of them might be invited to an interview at a later stage. Registering no opposition, he considered this a ‘green light’. Thanks to his active participation in group discussions he began building trust, which, like admittance, was facilitated by prior acquaintance with some of the UCG group members. The researcher-participant was thus quickly integrated into the group, and soon considered a ‘regular’ member. It is worth noting, however, that he was well aware that with chat apps, consensus cannot be assumed as a once-off ocurrence, as one cannot guarantee that every user (including future members!) has consented to the research, on account of the fluid aspect of interactions on WhatsApp (e.g. engagement can be intermittent, and not all users read all messages; new users can be added, as well leave a group). He did take care to send occasional reminders about his research, paying attention not to disrupt the conversation. Interestingly, amongst the informal commitments he made to the group members upon joining and disclosing his identity of researcher, was the promise to be accountable with respect to public presentations of the research results. Over time, group members felt he acted as their ‘voice’ and ‘representative’ within academia, and welcomed with enthusiasm information about, for example, public presentations of the research.

Digital ethnography allowed the researcher to observe from up close the interpretations, practices and everyday experiences that foster dynamic realities inside chat apps when it is conceived as a connection point between the online and the offline sphere of action (Boyd, 2008), or in a permanent transit between these two dimensions (Máximo, 2007). He could observe identity formation at the group level (the meso level of mobilisation), and, with the help of semi-structured interviews, explore the micro level of meaning-making, too. Broadcasting indignation and solidarity, the bottom-up character of UCG political action inaugurated a unique way for individuals to share stories, interacting with each other, and connecting locally in this novel online-offline civic engagement space. Exploring whether the ‘WhatsAppers’ represent a new form of political participation able to generate social change appropriating the affordances of the chat app, the research exposed how the UCG group was united under a leftist project formed by several social actors looking for a collective experience. He could identify the dynamics of discussion and coordination, while making a point about the validity of investigating private chats, instead of merely focusing on public groups (Garimella and Tyson, 2018; Resende et al., 2018). Users with similar political views came together but also recognised the right of other WhatsAppers with distinct opinions to present their views on the political situation. The group thus hosted also dissent and discussion, but within a basis of mutual respect. It evolved contingently over time, in response to the political situation it addressed, expressing demands, frustrations and projects to unite against the rise of the extreme right-wing faction and the beginning of Temer’s (2016-18) government. Although similar types of ‘connective action’ (Bennett and Segerberg, 2012) are frequently associated with fragmentation and atomisation, the research showed how private group chats on WhatsApp did help to foster creative interactions through high-speed engagement. It also showed dynamics of this nature can facilitate the emergence of a collective identity based on lower-common-denominators yet nonetheless useful to the cause (S. Milan, 2015). These identity elements, seemingly individualised and fluid, can be facilitated by active group administrators in WhatsApp acting as ‘soft’ leaders or ‘choreographers’ (Gerbaudo, 2012).

Towards an ethics checklist for research in (private) chat apps

Given the intimate nature of chat app interactions and the ‘false’ perception of secrecy that encryption brings about, ethical considerations are core prerequisites for engaging in digital ethnography in a way that is respectful of human subjects, especially when they are vulnerable individuals, as it is often the case with political activists.⁸ The field is inevitably slow at adjusting research ethics to upcoming ‘digital’ challenges. Yet such a research site and these methods are particularly prone to deception and misconduct. In this section we depart from the experience with UCG, to review existing reflections and provisions on research ethics and apply them to the field site of chat apps. In so doing, we aim at starting a much-needed conversation about the need and practice of research ethics in chat apps research, with a special concern for research on digital and/or media activism facilitated by chat app (and private/closed groups in particular). Our building blocks include documents prepared by professional organisations (e.g., Association of Internet Researchers, 2002, 2016), specialised groups (Ethics Feedback Panel for Networking and Security, 2014; Utrecht Data School, 2017), and existing contributions in the fields of media studies, humanities and the social sciences (e.g. Flacks, 2005; Pimienta, 2005; Chesters, 2012; Ess, 2013; S. Milan, 2014; Floridi and Taddeo, 2016; C. Milan and Milan, 2016; Zimmer and Kinder-Kurlanda, 2017; Ong, 2019). We also take into account the evolving legislative backdrop (e.g., the General Data Protection Regulation, which came into force in the European Union in May 2018 that has rapidly – even if questionably – become the new ‘gold standard’ of data protection worldwide) as well as the measures adopted by the industry following questionable data reuse practices such as those brought to light by the Cambridge Analytica scandal (see, e.g. Ward, 2018).

Generally speaking, researchers of all disciplines can prefer to avoid going through an ethics committee procedure, as these processes are tedious and complex, and often have the unwanted result of slowing down the research. What’s more, ethical considerations are only seldom made explicit in published manuscripts. To name just one, the field of social movement studies – one of the key disciplines devoted to the study of activism – has only seldom engaged openly with questions of ethics. This leaves a lot of colleagues thirsty for example and best practices.

In earlier works (2014), Milan identified four questions that pertain to the study of political activism and dissent, namely: relevance, which has to do with the nature of the researcher/research subject relationship, and asks what knowledge should be produced and for whom (see also Croteau, 2005; Flacks, 2005; Ryan, Salas-Wright, Anastario and Camara, 2010); risk, which interrogates the potential harms a research activity can entail for activists; power, which concerns the unbalanced relationship between research subjects and researchers (cf. C. Ryan and Jeffreys, 2008), and accountability towards research subjects, which includes also knowledge sharing. Milan and Milan later expanded the list into the STRAP framework, acronym for Sharing, Translation, Relevance, Accountability, and Power, designed to encourage researchers to engage communities on the ground as active agents of knowledge production (C. Milan and Milan, 2016). We now take these reflections one step forward to start addressing the lack of benchmarks and criteria for research on digital and/or media activism mediated by chat apps.

Conclusions

As the UCG case study of #UnidosContraOGolpe shows, activism today is increasingly facilitated by chat apps. New technological affordances, new practices of civic engagement, and new forms of surveillance and risks require novel approaches to research ethics. This article started from reflecting on the challenges of doing digital ethnography in a private chat group, to offer a number of pathways to help researchers to think though the limits of their research. To be sure, we do not want researchers to shy away from studying private chat groups. Rather, we intend to encourage academia to build strong practices of individual scholarship and community praxis in relation to research on chat apps, and digital ethnography in particular. We acknowledge ethical considerations by the research community to be a step forward from and a corrective measure against moral considerations by individuals. They are of paramount importance in fostering a healthy relationship between research subjects, especially when politically active, and researchers exploiting the affordances of digital technology to research ends. We believe that we cannot comfortably research media activism today without considering ethical challenges as well as the risk of surveillance and repression of activists on the ground.

We suggested a two-pronged approach. On the one hand, we should rethink and update ‘known’ ways of doing ethics, undertaking at least three conceptual operations: going back to the basics, positing as central the notion of ‘do not harm’, which allows to re-centre the research subject at the heart of the research process; avoid reducing research ethics to a one-stop checklist, to privilege instead a recursive, iterative and dialogic process able to engage research subjects; moving past the consent form as the sole and mere regulatory moment of the researcher-research subject relationship. On the other hand, while thinking through innovative ways of thinking ethics in chat app research, we ought to take infrastructure seriously, both the research site and the research ecosystem; embrace transparency and when the research question allows covert methods, avoid dishonest bypasses; and guarantee full anonymisation to research subjects to in order to future-proof findings and publications.

Far from providing a fully-fledged research agenda, this article hopes to serve as a way of kicking-off an urgent, collective conversation that can highlight ethical commitments to protect everyday citizens that appropriate chat apps for digital activism. Only in this way can academia implement best practices in exploring creative techniques to investigate media activism in chat apps today.

Competing Interests

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No 639379 – DATACTIVE, awarded to Stefania Milan as Principal Investigator; https://data-activism.net).

Author Information

Sérgio Barbosa is a PhD student in the program ‘Democracy in the Twenty-First Century’ linked to the Centre for Social Studies (CES) at the University of Coimbra, Portugal. In the academic year 2018-19, he has been an Erasmus+ exchange student at the University of Amsterdam, under the supervision of Dr Stefania Milan. His research explores emerging forms of political participation and resistance vis-à-vis the possibilities afforded by chat apps, with an emphasis on WhatsApp. Sérgio is a research associate of the DATACTIVE project and an active member of the Big Data from the South Initiative. He is also a member of the COST Action CA17135 ‘Constitution-Making and Deliberative Democracy’, and of Tecnopolitica, a ‘Latin’ research network connecting Brazil and Ecuador with Spain, Portugal and Italy. Contact: sergiosilva@ces.uc.pt.

Dr Stefania Milan (stefaniamilan.net), is a scholar of social movements specialising in new forms of political participation facilitated by technological innovation. She is Associate Professor in New Media and Digital Culture at the University of Amsterdam, and the Principal Investigator of two projects financed by the European Research Council exploring data- and algorithmic-mediated forms of civic engagement (see data-activism.net and algorithms.exposed). Passionate about building bridges between research, activism and policymaking, Stefania has been researching grassroots engagement with communication technologies since 2003. Her research has furthered the development of interdisciplinary theory in the realm of networked collective action, including on the ethics of research with vulnerable activists. She is the author of, among others, Social Movements and their Technologies: Wiring Social Change (Palgrave Macmillan, 2013/2016) and an editorial board member of the book series Critical Digital and Social Media Studies [University of Westminster Press]. Contact: s.milan@uva.nl.