Main theatre

Imagination gone rampant or Paranoia deserved? An Account of Folk Tales of Algorithmic Control among Creative Freelancers

Imagination gone rampant or Paranoia deserved? An Account of Folk Tales of Algorithmic Control among Creative Freelancers

Dr. Ana Alacovska, Copenhagen Business School
Dr. Eliane Bucher, Norwegian Business School BI
Dr. Christian Fieseler, Norwegian Business School BI
Division: Business

Short abstract

Just because you are paranoid doesn’t mean no one is out to get you
Joseph Heller, Catch-22

Increasingly, creative freelancers rely on online intermediary platforms such as fiverr or upwork for significant portions of their income. These platforms often govern their access to potential clients, decide on their visibility in ranked searches for their professional services, and impose reputation systems on them to highlight the best performers on the platform (Jarrahi and Sutherland, 2019). While not necessarily intended by the platforms themselves, for many of these freelancers, their work is marked by a radical uncertainty and opacity. In informal conversations, the assumed algorithmic nature of platform decision-making is a major contributor to this felt unease. At least in the imagination of many, algorithms play an increasing role in their market for labor, and these are in turn perceived as incomplete, partial and asymmetric information settings.
In our research outlined in this abstract, we are interested in these imaginations of the systems of algorithmic control, imaginations that are often set against the challenge of not knowing exactly what the algorithm does, if at all, and when it is at play. Against this background of freelance labor that suffers from being unable to meaningfully retrace decision-making, we describe a coping technique where users develop their own imaginations and theorization of the algorithm and its role in shaping the platform experience. In their ‘re-imagination’ of autonomous decision-making processes we describe in our contribution, the freelance users not only attribute agency but also motives and even personality to the algorithm.
We position our exploration and theoretical framing of this phenomena in the literature on organizational imagination and paranoia, where we are in particular informed by psychological mechanisms such as over-attention to negative happenings and attribution bias, paraidolia, the tendency to overdetect than to under detect agency, and a general tendency to humanize the system of algorithm control and agency, that is a general tendency to attribute agency to things users do not fully understand or grasp – in short, their hyper-active agency detection drive (cf. Bloch, 2016; Kramer, 2002).
Empirically, on the basis of 12’000 scraped comments from an online community of freelancers we show how communities perceive how they are being governed by the digital platform. Here, imaginations of the algorithm as an agent shaping their work and work prospects on the platform emerges as a recurring theme. We in addition collected visual metaphors from freelancers encapsulating how they perceive the algorithm to shape their experience on the platform (cf. Clarke and Holt, 2017). In subsequent interviews, building on their visual metaphors, we asked further how the (imagined) algorithm affected them and how they imagined different or better futures.

Figure 1 – Example of Visual Metaphors Collected


From both parts of the research, it became evident that in users’ ‘paranoid’ imagination of the algorithm (and its motives and rationales), remedial strategies were developed to ‘placate’ the imagined algorithm. To reclaim agency, many recontextualized instances which they did not fully understood, and gave a face and assumed behavior to their invisible nemesis, thus being better able to adapt their behavior for their future success (a technique which we describe as ‘pacifying the imaginary dragon’). In a wider sense, we propose, and illustrate for this context, folk theories as a promising lens to understand the observed behavioral adaptation, where individuals subjected to opaque governance systems show, based on such folk theories, efforts to make sense, as well as adapt and retake agency.
We see our contribution as an extension of theorizing on organizational autonomy and control (cf. Petriglieri, Ashford, and Wrzesniewski, 2018). Currently, models of platform work are bounded by theoretical discourse on autonomy (freedom, carrot, etc.) and control (forced responsiveness, stick etc.), where there is an identifiable (and real as well as material) source of control, that can be appeased (cf. Hamilton, Karahalios, Sandvig, and Eslami, 2014). We would extend that towards a more constructive view, where individuals fundamentally fail or are at least challenged to understand how to deal with this change in the locus of control, how to deal with the inscrutability of algorithmic management, and propose to delve deeper into paranoidstructural aspects and quasi-religious ‘sense-making’, as well as conspiracy and folk-theorizing as a contribution to explaining new ways of inscrutable control.


Bloch, C. (2016). Passion and paranoia: Emotions and the culture of emotion in academia. London: Routledge.
Kramer, R. M. (2002). When paranoia makes sense. Harvard Business Review, 80(7), 62-9.
Hamilton, K., Karahalios, K., Sandvig, C., & Eslami, M. (2014, April). A path to understanding the effects of algorithm awareness. In CHI’14 Extended Abstracts on Human Factors in
Computing Systems (pp. 631-642). ACM.
Petriglieri, G., Ashford, S. J., & Wrzesniewski, A. (2018). Agony and ecstasy in the gig economy: Cultivating holding environments for precarious and personalized work identities. Administrative Science Quarterly, 0001839218759646.
Clarke, J., & Holt, R. (2017). Imagery of ad-venture: Understanding entrepreneurial identity through metaphor and drawing. Journal of Business Venturing, 32(5), 476-497.
Jarrahi, M. H., & Sutherland, W. (2019, March). Algorithmic Management and Algorithmic Competencies: Understanding and Appropriating Algorithms in Gig Work. In International Conference on Information (pp. 578-589). Springer, Cham.