Since then, these questions and more have been buzzing in the back of my mind as I revisit my Work Plan and timelines to account for project delays because of COVID-19. I’ve been fortunate because I designed my study in a way that avoids front-loading empirical tasks in Work Package 1 (WP1). The beginning stages of the project unfold somewhat slowly and revolve around information collection and curation which can be done remotely. My main objective for WP1 overall comprises horizon mapping of AIEd discourse and initiatives in Ghana and South Africa; identification of recurring discursive themes; developing a better understanding of the areas under study; and refining frameworks and instruments to be used for data collection. I am on my way to completing much of this as I also move through organising recruitment of two postdoctoral researchers.
To be honest, much of my attention so far has been fixed on organising the technical and administrative details of the project, such as procurement, vires to account for complications related to the pandemic, and understanding legal concepts specific to contracts drawn between institutions. Because the project is a multi-country study, there have been new things I’ve had to learn pertaining to country-specific tax laws; risk assessments; agreements across universities to host postdoctoral researchers; recruitment; and of course, ethics approvals to conduct research in schools. Trying to project manage only remotely and during a pandemic has been quite the experience.
I have learned to not underestimate the amount of time it will take to work through administrative and legal details. These requirements can be easy to miss (and on occasion misunderstand) as they are sometimes buried under layers of other information, more documents, redirects, and – at times – broken links. Things I’ve learned the hard way:
- Bureaucracy happens, resistance is futile;
- You will make mistakes, but you can learn from them;
- Computer systems break down when you need them the most.
Off the record for the moment, please and thank you: The older I get, the more I see, the more I am sure #3 is cross-cultural.
While I write this journal entry, I continue my search for schools in South Africa as potential sites for ethnographic focus. I have been reading about AI/digital initiatives in three provinces in the hopes of identifying schools to participate in the project. I will write more on this in a separate entry. For now, I have it on good authority that schools in the following provinces may offer helpful comparisons of the different paths taken to enacting inclusive, accessible, and scalable technological innovations in education:
- Western Cape
- Gauteng Province
- East End Cape
I am also searching for comparable schools in Ghana. I have been provided with enormous help from the University of Cape Coast as I try to make sense of the digital education terrain in the Ghanaian context. I am grateful for this.
While the project kick-off meeting was initially focused on important administrative processes, we eventually turned to theoretical approaches, ideas, policies, and other sources of information that might offer a way forward as I grapple with putting together an ecological framework for studying AI in education and development.
A Theory of Niche Construction
A much appreciated benefit of this fellowship is that I have the time to pause, read closely, and reflect on ideas I encounter while I learn more about areas relevant to the present study. As I think about various ways of collecting, approaching, and presenting my research, I am also asking myself the following: What is the problem I am trying to understand? What is the best methodological framework for answering my research questions? … Why an ecological approach? Reflecting on these questions and exploring the internet brought me to Richard Dawkins (1982)[i] and the concept of niche construction (evolutionary biology). I am trying to figure out the challenges I might face moving Niche Construction Theory to Sociology of Education. Suggestions for further reading are welcome, as always.
Relatedly, I’ve long been fascinated with the work of Jakob von Uexküll (1864 – 1944), a biologist and bio-semiotician interested in animal behaviour and the semiotics of life. Uexküll’s most well-known contribution to research is the concept Umwelt. He argued that all animals, despite levels of complexity, must be understood as subjects inhabiting environments which are constituted and made meaningful through particular ways of perceiving and acting. Uexküll referred to these environments or worlds as Umwelt (literally translated: “surrounding world”).[ii] I began thinking about and applying concepts of world and niche in my research on biosurveillance. I wonder if it is useful for me to continue working on these heuristics for my present study on education, technology, and environment?
I’m hesitant to take up too much more of your time, but I’d like to share something else I found helpful. I recently came across a paper about ecologising sociology by Jonathan Murdoch (2001).[iii] The paper gave me a general sense of the challenges of doing sociology of education research using an ecology framework. Murdoch invites us to think about how we might apply Bruno Latour’s reflections on ecology as a way of focusing on not simply the social but also the environmental.[iv] In articulating an ecological approach, Murdoch points out that Latour was attempting to reimagine the role of sociological analysis from the framework of Actor-Network-Theory (ANT) and extending this work on science into a “general theory of the relationship between the two ‘Great Domains’ of nature and society” (p. 117). On this view, the social and environmental exist in an interconnected relationship.
Latour proposes that ANT is a useful analytic framework for studying ecological phenomena. He maintains, however, that separating constituent elements of environmental/social issues in line with the ontological categories of society and nature do not make much sense for analytic purposes. He suggests that the distinctions made between human and non-human animals are anthropocentric, arbitrary, and privilege a particular kind of entity. To properly appreciate the various dimensions of ecological phenomena, he argues, one requires an approach that situates human beings in a complex dynamic of heterogenous relations which also requires rethinking arbitrary ontological distinctions between society and nature.
As I read more, I am mindful of the criticisms levelled at this kind of mix of ANT and ecology. Unless I am misunderstanding, one main critique seems to rest on whether or not the blurring of distinctions between nature and society is helpful when dealing with issues of environmental accountability. As scholars have noted, if we wish to take environmental rights and protections seriously when doing research (e.g., sociology of education), we cannot afford to run into the problem of wiping away distinctions made between human and non-human animals/entities as a move towards redefining value and place in the moral community. This is a trap that Latour acknowledges he falls into when he suggests doing away with such ontological distinctions: If we accept this blurring of distinctions, we must also accept that humans become equally not-culpable as their non-human animal counterparts, for example, for harms to the environment. If I understand this correctly, equalising ontologies (is this even a process?) between humans and non-humans means that we – as human beings – are also released of our moral obligation to be held fully accountable for our impacts on the natural environment – same as non-human animals. If I am to apply this kind of ecological perspective to sociology of digital education, how would I do it and what other conceptual problems might I encounter along the way?
AI in education: Whose ethics?
I’ve also started reading more widely about fairness. Based on what I’ve found so far, general conceptions of fairness[v] in AI/ML tend to be informed by factors that can be quantified. Discussions of fairness in the ML literature, for instance, articulate fairness as a process that can be understood, judged, and acted upon by a machine, by algorithms. However, there are other dimensions to fairness we ought to be taking into account as writers have pointed out.[vi] Some of these include aspects of fairness that are more difficult to capture through and as a result of an algorithmic process. With this in mind, can fairness be universalised or is it only meaningful when it is relative to context-specific factors? How might we design fairness-enhancing interventions in AIEd systems that account for context? What does fair to the environment mean and look like in educational technology practice?
As I read more about fairness (and I have a long way to go), I am also trying to untangle what artificial intelligence refers to in popular discourse and academic research, but this is a story I will save for another day.
Till next time.
- Darwin, C. (1869). On The Origin of Species. The Project Gutenberg. Retrieved: https://www.gutenberg.org/files/1228/1228-h/1228-h.htm
- Mbembe, A. (2001). On the Postcolony. Berkeley, CA: University of California Press.
- Singer, P. (2016). Famine, Affluence and Morality. Oxford, UK: Oxford University Press.
- Wynter, S. (1996). Is development a purely empirical concept or also teleological? A Perspective from “We the Underdeveloped.” In Yansané, A.Y., Prospects for Recovery and Sustainable Development in Africa. Westport, CT: Greenwood Press, pp. 299-316.
[i] Dawkins, R. (1982). The Extended Phenotype. Oxford, UK: Oxford University Press.
[ii] Schroer, S.A. (2019): Jakob von Uexküll: The concept of Umwelt and its potentials for an anthropology beyond the human. Ethnos. doi: 10.1080/00141844.2019.1606841
[iii] Murdoch, J. (2001). Ecologising sociology: Actor-Network Theory, co-construction and the problem of human exemptionalism. Sociology, 35(1),111-133. doi:10.1177/0038038501035001008
[iv] Latour, B. (2005). Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford, UK: Oxford University Press.
[v] Friedler, S.A., Scheidegger, C., Venkatasubramanian, S., Choudhary, S., Hamilton, E.P. & Roth, D. (2018). A comparative study of fairness-enhancing interventions in machine learning. Retrieved: https://arxiv.org/abs/1802.04422
[vi] Greene, D., Hoffmann, A., & Stark, L. (2019). Better, nicer, clearer, fairer: A critical assessment of the movement for ethical artificial intelligence and machine learning. Proceedings of the 52nd Hawaii International Conference on System Sciences. HICSS. Retrieved: https://bit.ly/30ra7Uu