Why a Humanist Ethics of Datafication Can’t Survive a Posthuman World

Geoffrey Rockwell and Bettina Berendt's (2017) article calls for ethical consideration around big data and digital archive, asking us to re-consider whether information wants to be free. In outlining how digital archives and algorithms structure potential relationships with the other whose testimony has been digitized, Rockwell and Berendt highlight how data practices change the relationship between research and researched. They make a provocative and important argument: datafication and open access should, in certain cases, be resisted. They champion the careful curation of data rather than large-scale collection of big and dirty data, pointing to the ways in which these data are used to construct knowledge about and fundamentally limit the agency of the research subject by controlling the narratives told about them. Rockwell and Berendt, drawing on Aboriginal Knowledge (AK) frameworks, amongst others, argue that some knowledge is just not meant to be openly shared: information is not an inherent good, and access to information must be earned instead. This approach was prompted, in part, by their own work scraping #gamergate Twitter feeds and the ways in which these data could be used to speak for others, in real time, without their consent.

From our vantage point, Rockwell and Berendt's renewed call for an ethics of datafication is a timely one, as we are mired in media reports related to social media surveillance, electoral tampering, fake news and micro-targeting on one side. Thanks, Facebook. On the other side, academics fight for the right to collect and access big data in order to reveal how gender and racial discrimination are embedded in the algorithms that structure everything from online real estate listings, to loan interest rates, to job postings (American Civil Liberties Union 2018). As surveillance studies scholars, we deeply appreciate how Rockwell and Berendt take a novel approach: they turn to a discussion of Freedom of Information (FOI), Freedom of Expression (FOE), Free and Open Source software, and Access to Information. In doing so, they unpack the assumptions commonly held by librarians, digital humanists and academics in general, to show that accumulation and datafication is not an inherent good. This is coupled with a healthy skepticism surrounding techno-utopic big data claims. Freedom and sharing economy rhetorics aside, it is quite evident that datafication does not benefit all equally, but favors those who already hold power, increasing the gap between 'haves' and 'have nots' (Srnicek 2016), and, following Virginia Eubanks (2018), automating inequality.

One of Rockwell and Berendt's key points is that the act of digitizing and mounting digital archives can silence others who have some claim to the culture digitized: data speak for others without their consent (and often without their knowledge). In this way data are placed on equal ground as testimony, as narrative. Therefore, an ethics of datafication should be rooted in a relationship of care for the other that generated these data, and the infrastructures of digital archives should be constructed in a way that might limit access to data or the stories that might be told through them. In some cases, an ethical practice on the part of digital humanists might be to avoid datafication entirely. In response to this line of argument, we'd like to offer a few provocations of our own. The notion that digitization may expose private or even sacred narratives to those outside of culture-bound communities presumes a narrative quality to digital information that it does not inherently possess. In order words, we regard narrative as a human construct rooted in interpretive hermeneutics that must enter into a co-productive relationship with digital information in order for narrative to emerge as its avatar. As such, narrative may emerge from any data set¬-whether complete or incomplete-and present an interpretive conclusion. This introduces new ethical questions regarding the production of informed narrative versus those that might be misleading or predicated on incomplete information. Thus, our first counterpoint to Rockwell and Berendt's article is rooted in the pragmatics of declining datafication in an era of surveillance capitalism (Zuboff 2016).

1. Archives and Infrastructure: The Epistemological Consequences of Declining Datafication.

Although nearly every living person in the developed world is subject to the forces of information extraction and, as a corollary, to datification, these processes are not collective efforts. While the early internet promised a democratic new territory for users to explore, exchange information, and interact across a diverse range of small, independent sites and locals, the internet-as we know it today-has matured beyond its promising beginnings, mutating into something altogether different (Tufekci 2016). Indeed, as the internet shed its democratic origins (if they ever truly existed), network traffic became increasingly concentrated around several key sites or platforms that now command the lion's share of bandwidth allocation and surveillant capacity. As such, the ability to extract, parse, and make sense of user data is largely concentrated in the hands of a few powerful information brokers who have a vested interest in engineering the public and making these more tractable, empirical non-human constructs speak (Tufekci 2014). Indeed, even national programs directed at the surveillance of online data for the maintenance of international intellectual property rights, state governance, and for national security purposes rely heavily on these brokers to guarantee a continuous stream of information (See: Lyon 2015).

In the face of such practices, an ethical stance that actively limits datification and/or the recombination of data to tell new narratives inevitably demands careful, considered, and costly construction of data infrastructures on the part of researchers. Yet Rockwell and Berendt's curated data edifices, while ethical and respectful to the generators of said data, would be vastly overshadowed by competing corporate infrastructures, which are built with no such restrictions. If the public sector, including academia, fails to keep pace with digitization and information extraction efforts, this ultimately increases the asymmetrical imbalance between private sector and corporate information brokers and the general public. The narratives that data tell will be predominantly constructed by corporate actors, with material interests. And so, while an ethics of datafication is important, it must also be accompanied by efforts to regulate and make more transparent that corporate data practices operating in parallel to scholarly archives.

Although Rockwell and Berendt question and problematize the potential of big data to advance human knowledge, its true ability to generate valid insight need not be placed into question here. If the archives and analytical toolkits that constitute big data are possessed, produced, and manipulated by only private or government agencies, there can be no independent verification or public inquiry into any inference that they might make. As such, any increase in the asymmetry of the possession of digital information translates directly to an asymmetry in the ability to make unverifiable truth claims-providing, in essence, an epistemic carte blanche to corporate or state actors. We consider this potential for corporate or state epistemic hegemony to be a major cause for concern in the formulation of Rockwell and Berendt's ethical position.

More directly, in addressing Rockwell and Berendt's concerns that secret indigenous mythologies or histories might be made-through the process of digitization-available to audiences outside of those to whom they rightfully belong, we counter that such digitization is already occurring with or without the blessing of public sector actors, academics, or private citizens, and is traded, popularized, erased, commodified, or otherwise adapted for the purposes of either monetization or government profiling. An ethical stance rooted in declining datafication thus could generate further inequalities rather than resolve them.

2. Whose Story Do Data Tell? Privacy, Privilege and Posthuman Ethics

Our second provocation is that the ontological perspective from which Rockwell and Berendt develop their ethical critique relies on a humanistic framework that the processes of informatization and digital surveillance actively cannibalize (and may have already undone to a significant degree). The emergence of big data, mass digitization, digital surveillance, information extraction, and the rise of the so-called internet of things have prompted an entire posthuman literature that sought (and seeks) to de-centre the human figure from its privileged position in political ontology. As such, we find an ethics rooted in a humanistic framework to be potentially anachronistic. Here, we argue against a humanistic ethics of digitization along two axes. The first applies pressure to the notion of a digital privacy as a human right. Like all rights, digital privacy is unevenly administrated and accessed. Privacy is made manifest by neither law nor sentiment; it must be enacted, protected, and enforced. In the case of digital privacy, the ability for one to control whom they share information with and how it is used, is deeply stratified.

For those individuals who find themselves under an increased level of digital surveillance or information extraction, the ability to withdraw from the production of archive is limited. Police records, social assistance records, immigration forms, health records, etc.-themselves crosscut by, and prone to, reproducing, histories of racism, classism, cissexism, and homophobia- will continue to be entered into the archival record. Indeed, it is true that certain factions may be able to avoid surveillance, information extraction, and the digitization of their life's story, and yet we see this kind of digital or surveillant abstinence as coming packaged alongside a difficult ethical problem: it creates a situation wherein the positive aspects of abstaining from datafication are bestowed upon those privileged enough to escape continuous legal, medical, epidemiological, financial, and moralizing government surveillance, and are able to contend with-or even benefit from-an archive composed of materials that reflect and recapitulate unjust, discriminatory premises and practices. Taken alongside the way that powerful data brokers might increase their stranglehold on big data archives and the algorithms through which they are made to speak truth, such a prospect should frighten us.

While Rockwell and Berendt actively resist framing data in agential terms (e.g. critiquing the idea that information wants anything, let alone to be free), they frame data as indexical - linking them to individualized narratives (e.g. datafication is an appropriation of voice). In this move, Rockwell and Berendt risk over-simplifying what data are. Our point here is this: our corpus of data only rarely point back to human actors, and even more rarely do they capture authorial or spoken intent. For the most part, the stories data tell are not about us, and thus an ethical approach to datafication that is rooted in a humanistic framework may be obsolete already.

A posthuman world is also one that is post-anthropocentric. Although much environmentalist and ecological literature has sought to de-privilege the human animal as the sole beneficiary of our planetary resources, only recently have they intersected with discussion centred around digitization and big data. In Benjamin Bratton's The Stack: On Software and Sovereignty (2016), Bratton conceives of an accidental megastructure of both organic and inorganic components drawn together through the omnipresence of machine interfaces, universal addressing systems, software, hardware, and various sensory machines. In the Stack, Bratton draws little distinction between the human and non-human, deeming them both user inside the technological apparatus that has engulfed not only the surface, but the geological strata of our earth.

If we focus only on those ethical concerns that arise from a humanistic vantage point, we ignore an equally provocative and prudent ethics that seeks to protect the unhuman and to give it voice in a moral calculus. It is in the vein of Donna Haraway's recent writing in Keeping with the Trouble: Making Kin in the Cthulucene (2016) and Patricia MacCormack's Posthuman Ethics (2012) that we render these critiques. To privilege the human need for privacy in these posthuman times is to deny adequate consideration to the ways in which our data-even our compromising data-may contribute to the healing of our damaged planet. Moreover, data gathered from internet-connected non-human sources (already outnumbering human sources of data), must also skew our understanding of the ethics of digitization. For the first time, our oceans, rocks, minerals, flora, and fauna, may be able to enter their own histories and narratives into the archive. While digitization may indeed pose ethical problems for humans, the broader ethical questions that circumambulate digitization and information extraction writ large reach far deeper and much farther out than those that elect (rather arbitrarily) humanist principles as their central dogma.

References

American Civil Liberties Union. 2018. Judge Allows ACLU Case Challenging Law Preventing Studies on 'Big Data' Discrimination to Proceed. American Civil Liberties Union. April 2, 2018. https://www.aclu.org/news/judge-allows-aclu-case-challenging-law-preventing-studies-big-data-discrimination-proceed.

Bratton, Benjamin H. 2016. The Stack: On Software and Sovereignty. Cambridge, MA: The MIT Press.

Eubanks, Virginia. 2018. Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor. New York, NY: St. Martin's Press.

Haraway, Donna J. 2016. Staying with the Trouble: Making Kin in the Chthulucene. Durham, NC: Duke Univiversity Press.

Lyon, David. 2015. Surveillance After Snowden. Cambridge; Malden, MA: Polity.

MacCormack, Patricia. 2012. Posthuman Ethics: Embodiment and Cultural Theory. Farnham, Surrey, England: Routledge.

Rockwell, Geoffrey, and Bettina Berendt. 2017. Information Wants to Be Free, Or Does It?: The Ethics of Datafication. Electronic Book Review 2017 (December). https://doi.org/10.7939/DVN/10253.

Srnicek, Nick. 2016. Platform Capitalism. Cambridge, UK: Polity.

Tufekci, Zeynep. 2014. Engineering the Public: Big Data, Surveillance and Computational Politics. First Monday 19 (7). http://firstmonday.org/ojs/index.php/fm/article/view/4901.

---. 2016. As the Pirates Become CEOs: The Closing of the Open Internet. Daedalus 145 (1): 65-78. https://doi.org/10.1162/DAED_a_00366.

Zuboff, Shoshana. 2016. Google as a Fortune Teller: The Secrets of Surveillance Capitalism. Frankfurter Allgemeine Zeitung, March 5, 2016. http://www.faz.net/aktuell/feuilleton/debatten/the-digital-debate/shoshana-zuboff-secrets-of-surveillance-capitalism-14103616.html.

Cite this Essay: