Some thoughts about the automation of educational labour

Work automation involves two different scenarios. The first scenario entails the development software or hardware systems that can augment social practices; the second scenario is based on the creation of autonomous, self-organising systems that can completely supplant humans in a particular line of work. This distinction has been captured effectively by Aaron Benanav: ‘with labour-augmenting technologies, a given job category will continue to exist, but each worker in that category will be more productive. By contrast […] no matter how much production might increase, another telephone-switchboard operator or hand-manipulator of rolled steel will never be hired’.

Now, it is important to remember that not all tasks can be automated, and indeed there is a correspondence between the nature of work in large labour-absorbing sectors and the lack of automation. For instance, automation has not impacted in any significant way on textile work (sewing) and, notably, on first-link electronic assembling, which occurs before electronics are sent further up the productivity chain towards more ‘advanced’ semi-automated factories.

Applied to education, this line of reasoning has two consequences.

Firstly, the magnitude of the demand for teaching as a form of work is the first factor to consider when speculating on the future of automation in education: the higher the demand, the less automation will be viewed as a viable proposition, because societies benefit greatly from sectors that can absorb human labour. Employed humans, however inefficient or hard to govern they may be, produce healthy economies.

Secondly , there is the nature of pedagogical practice, which cannot be fully automated because too much of it is relational and embodied – a ‘form of life’ and an adaptive component of the human experience, manifested in multiple forms during the life course, sustained by an evolutionary and biological substratum (humans are not the only complex mammals who teach their young) but also deeply embedded in linguistic and cultural traditions: a sociocultural activity.

Once these two reasons are accounted for, what is left is a view of automation as cybernetic governance – a form of control that does not pursue human replacement, but standardisation, docility, and the stultification of practice. While task automation reduces human activity in some ‘core’ areas, it generates new trivial tasks that demand people to coordinate effectively with a plethora of platforms and data-based administrative systems.

Thus, the true horizon of automation in education becomes apparent – not lights out automation, but the apprehension and control of educational practice and leadership in the name of managerial accountability. Not robots in the classroom, but teachers acting in standardised and predictable ways, unable to operate autonomously when unplugged from the digital infrastructure.