Digital test beds as sites of shifting targets and scholarly interventions

Abstract

This commentary is a response to Thao Phan’s article that tracks the relationship between contemporary digital test beds and histories of colonialism, racism and empire. Building on Phan's arguments, I suggest two possible further expansions: one is to link drone experiments with global logistics and militarism, and track how the end of experiments enables a slippage between military and commercial targets of emerging technologies by iteratively producing new forms of targets. The other one is to ask how testbeds could also operate as sites for scholarly experiments and interventions on the test, to different ends and to a different success.

Digital technology test beds appear everywhere—in towns, rural areas, policies, organizations, and everyday life. Thao Phan asks what makes people and places ideal technology test sites and subjects, and urges scholars to consider how technology testing “in the wild” emerges from particular conjunctures of power, knowledge, and histories of colonialism. By tracking Australia's position as a historical and contemporary site of technical experimentation, Phan explores how the logic of the test bed links contemporary drone delivery tests by a Google subsidiary firm with Britain's devastating nuclear weapon testing in the 1950s and with past experiments of colonial medical science that tested the limits of whiteness. Through these linkages, Phan shows how “ideal” test sites and subjects are made through projections, imaginations, and technocratic management practices at the nexus of knowledge and violence, power and subjugation, racism and diversity, global aspirations and nation-building projects, now and in the past. Through such a discussion, Phan makes a substantial and timely contribution to ongoing debates on digital test beds that ground them in histories of colonialism, knowledge, and empire while carefully uncovering the qualitative differences between past and developments. The article focuses on Australia, and provides many inspirations for further thinking about the politics and consequences of digital technology testing. On the basis of Phan's account and propositions, I reflect on the question of what other possibilities such tests generate (Marres and Stark, 2020) and how technology test beds become sites of shifting targets and arenas for scholarly interventions.

From testing to targeting

The ongoing experiments with drone delivery technologies that Phan discusses are fundamentally about experimenting with forms of logistics. Contemporary logistics as a business practice evolved from military practice and shares a common history with militarism (Cowen, 2014). Therefore, the particular case of drone delivery testing that Phan describes invites us to consider how military techniques for targeting (Parks and Kaplan, 2017) and controlling subjects through drones migrate into the commercial domain. Such a consideration might fruitfully embrace the notion of vertical mediation that media scholar Lisa Parks proposed to consider how the use of aero-orbital technologies and spaces to construct, reconstruct, and destroy lifeworlds from above reshape the “material conditions for individual and social bodies, produce and circulate views of life on earth, and assert new global hegemonies” (Parks, 2018: 9). From such a perspective, an obvious question to pose here is how test subjects of drone delivery become not only subjects of tests but also commercial targets through the different forms of surveillance, knowledge, and control enabled by drone delivery. How does this type of tech experiment expand the pervasive digital logics of datafication, monetization, and personalization to enable commercial targeting through the vertical domain? What forms and configurations of labor produce these emergent forms of targeting? Companies such as Google already have technologies that produce a macro view from above through systems such as Google Earth, but these views lack the granularity and data that create a complementary, much more intimate relation to individuals in their homes. Drone delivery tests offer the promise of enabling the combination of macro and micro views from above, expanding forms of social and market controls through the vertical domain (see also Andrejevic and Volcic, 2023). Therefore, such a perspective invites an expansion of Phan's argument beyond tracking the entanglement of testing with projects of colonialism to additionally trace its linkages to militarism and logistics and how the blurring of lines between tests and the many forms of targeting occurs via drone technologies.

To uncover these linkages would require its own study, but in Phan's work, we can already see some of the logics through which such slippage between different domains, tests, and targets could happen. A point that Phan emphasizes through the case of experiments is how they evolve through failure. When people complain about noisy drones or when the technology malfunctions, their experiences do not suspend or end the experiment but rather prompt the development of new versions of the drones that would be more sensorially attuned to their social settings. Phan discusses how this form of response to failures instantiates a mode of “endless versioning” (Halpern and Mitchell, 2022) that normalizes experiments and helps technology developers make the case for the universal transferability of technologies to other social settings. This mode of endless versioning is what eventually takes the experiment off its initial form and regime of oversight and turns it into a site of prototyping that not only tests but also develops new targets and relations between test subjects and testers, eventually becoming a hack rather than a test. To hack means, among other things, to break up soil and ground as part of processes of cultivation. Through endless versioning, new test subjects and targets for tests are perpetually cultivated in the test bed.

An example from a test bed in Sweden illustrates this point. Several years ago, I was involved in a research project to study frictions that emerged through test beds for smart electricity grids. Our research group examined an experiment that tested a technology developed by a Swedish tech start-up on 300 households in the province of Uppland, north of Stockholm. The start-up developed an app, a set of sensors, and data-driven decision-making software that was later called “AI.” These devices were inserted in homes with the intent to track people's shower patterns, use of rooms, and routines in their homes, with a focus on their energy use, and then to seamlessly control their electricity-driven heating system to tune it a degree up or down for the benefit of the grid operator, who wanted to reduce energy consumption overall in the region in peak times to save on costs and its use of oil. The test trialed the idea of “seamless” control and the efficacy of the algorithm. Similarly to Phan, we noted failures when homes got overheated or froze, or when people reported on conflicts with family members over the optimal temperature of the home that prompted them to tweak or turn off their test devices (Velkova et al., 2022).

However, we also found that the devices in the initial 300 experimental homes and the algorithm that steered them remained in use by the start-up long after the official end of the experiment. The start-up enrolled them in new tests that focused on other qualities and possibilities of the system and in novel constellations of test subjects who were assembled across multiple test sites, past and new (see a discussion of one such example in Rohracher et al. (2025)). In parallel, the start-up worked to commercialize its products and create new services for a general market of homeowners. In this mode of endless versioning of the algorithm, app, and sensors, the relationship between the actors’ roles in the test evolved as the test subjects eventually became targets of commercial logics and new kinds of experiments with the devices. This example shows how digital technology test beds enact a slippage between test subjects and targets of commerce while exploiting the governing logic of the experiment as a trial that appears to have a clear start, an end, and a resolution but never ends in practice. As the test continues to yield new knowledge, it also triggers new relationships and opportunities in the emergent gray zone space of the test, where technology owners have the power to shift logics, extract value, and continue trialing and developing their technologies long after the test ended by leaving the test equipment in place. In this conjuncture, test subjects might voluntarily enroll in the experiment, thinking that they could contribute to the common good (e.g. how the energy grid could be managed more effectively and thus sustainably), only to become involuntarily objects of commercial targeting though the same tests and actors who set the experiment up. In this sense, experiments are hacks of regulatory systems and social life, as they creep into novel domains and create new targets through the logic of endless versioning that iteratively redefines the test targets and subjects.

Scholars testing the test

The slippage between the test subjects and commercial targets, both in Phan's example and the example from our research in Sweden, begs the question about the interventions that scholars could or should make in the space of such test beds. For instance, the smart energy experiments in Sweden generated many questions among our research team members about the ethics of repurposing the test, the reuse of test subjects’ houses and systems without their knowledge, the strategic cultivation of non-knowledge in the context of the experiment in general, and the reuse of data amassed through the experiment toward making multisided markets that included energy infrastructure management and commercial development. We wondered what to do with these questions. Could we have redefined the test? Could we have interrupted its logic? What could we do with our knowledge of these issues?

In an earlier conversation that unfolded on the pages of this journal, Minna Ruckenstein (2025) called for scholars to take positions as troublemakers and develop different trials to provoke engineers, and to foreground differences in perspectives, value conflicts, as well as differing approaches to sociotechnical issues. In our experiment, we eventually tested this stance and the testers. We asked the testers to reflect on the ethics of their never-ending experiment and remaking of test subjects into targets of multiple subsequent tests and multisided markets. The test leaders listened to us and admitted that they could have done things differently but did not do anything to change the situation—they were interested in the test of the technology itself and its commercial and infrastructural potentials, but not in its social and ethical aspects. They also valued being accountable to their funders more than to their test subjects, who lost their importance once they agreed to be part of the technical test. This stance prompted us to evaluate the forms of troublemaking that could be created in the space of this particular experiment and constellation of actors. Eventually, we had to shift attention to other matters when our funding ended and the academic structural logics pushed us to refocus elsewhere. Our trial to be troublemakers failed and showed how our efforts were also shaped and constrained by our own academic economies of time and logic, inventiveness, and position. This does not mean that we should give up testing technologists and making trouble, but underscores the need for inventiveness, persistence, and continued involvement; to remain with the experiment and the troubles it creates long enough to trigger change and engagement; and to create and maintain pockets of possibility that could make more socially sensitive futures with emergent technologies.

Our experience makes me wonder what a test on the testers could look like in the case of Phan's drone delivery in Australia? Who could be the target and in what ways? What responses would the test generate? Henceforth, moving beyond Sweden or Australia in the turbulent world of global geopolitics and ongoing expansion of global militarism, not least through drone deliveries that carry bombs rather than packages, Phan's case leaves open an obvious intervention for all of us: to pursue further inquiry on the ways in which test beds in different places around the world enable the shift between the military and commercial uses of emergent technologies, and track the endings of experiments, knowledge-power relations, and values that enable this shift.