A new report from Alexandra Mateescu and Madeleine Elish draws attention to the considerable embodied and material work that takes place around the introduction of data-driven and automated systems into social contexts. As they put it, the report is concerned with …
“the ways in which automated and AI technologies tend to mask the human labour that allows them to be fully integrated into a social context while profoundly changing the conditions and quality of labour that is at stake” (p.4)
Using the example of data-driven farming, Mateescu and Elish discuss the material and cultural implications of new systems being used by specific people within existing norms and practices. These descriptions have clear parallels with the implementation of similar technologies in schools.
For example, data-driven systems might require new work routines and changes to physical infrastructure. This might include reorganizing the layout of the classroom to facilitate optimal sensor readings, or installing window shades to ensure more complete camera visibility.
These technologies also imply cultural shifts in the operational and pedagogical logics of the school. For instance, a group of students or a body of curriculum knowledge “must now be conceptualized as a complex dataset to be managed through other digital information and digital tools”
In addition, Mateescu and Elish also raise the issue of how automated and AI technologies are not simply ‘deployed’ into settings such as schools – rather, the introduction of a new system or machine requires an extensive ‘human infrastructure’ to mediate, moderate and troubleshoot its demands and quirks.
To illustrate this point, Mateescu and Elish use the example of supermarket workers supporting the introduction of ‘self-service checkouts’ in their stores. In particular, they highlight how work is reconfigured, as workers are forced to take on new roles as guides and interpreters for confused end-users, or as trouble-shooters to provide a quick fix for the inevitable teething-troubles and breakdowns.
Again, there are clear parallels to explore here in terms of schools. For example, working to smooth the integration of any data-driven automated systems with the pre-existing routines and practices of a school is not easy. Mateescu and Elish point to the “tricky and sometimes counterintuitive” labour that is required to get these technologies up-and-running.
If all goes well then this might be described as a process of ‘harmonisation’ – what the authors refer to as ‘smoothing out the rough edges of a technology’. However, in reality it is likely to involve an ongoing process of disruptive work-arounds, awkward bodges and temporary solutions – many of which require compromises on the part of the school and the people already working there.
A key issue that arises through these examples are the individual risks involved in being involved in these processes. The flipside of being an ‘early adopter’ or ‘guinea pig’ is being exposed to the unintended consequences, glitches and bugs. The dominant model of technology development relies on problems being worked out on the ground. Yet this means that teachers’ and students’ ‘performance’ is likely to suffer as a result.
Mateescu and Elish argue that these risks (and costs) are disproportionately borne by vulnerable/precarious groups rather than those who are implementing the systems. Certainly, it will not matter that much to an LMS provider if a class does not get full access to online readings while their school is still ‘finding its feet’ with the system. However, this glitch does matter very much to the children concerned (who might only get one shot at taking that class), or the teacher (whose performance metrics are likely to suffer as a result). Schools are not places where individuals can afford to bear the consequences of relying on systems and processes that are in a perpetual ‘Beta’ state.
All told, Mateescu and Elish draw attention to a range of important factors to take into consideration when thinking about the datafied school. In particular, we need to properly consider how a range of people within schools have to work hard to accommodate these technological incursions … all of which involves the reconfiguration of “ingrained community norms, interpersonal relationships, daily routines, and skill sets” (p.6). This raises a number of interesting questions for our own school-based ethnographic studies of data-in-context …
- How is the continual adoption and adaption of data-driven systems in a school altering physical infrastructure, processes, and norms within that school – what material and cultural accommodations are being made?
- What work is required to accommodate and integrate data-driven systems in a school?
- Who is doing this work, and how is it rendered visible and/or talked about?
- Once a data-driven system is relatively stable, what are its everyday effects on work responsibilities and conditions within a school?