Existing methodologies in technology ethics predominantly take individual technologies, applications, or artefacts as their primary unit of analysis, treating technology as a relatively well-defined and discrete object of evaluation.
Recent philosophical work on explainable AI has explored mechanistic methods targeting the causal structure of deep neural networks. However, the ever-growing complexity of deep learning models challenges purely decompositional strategies.
In their 2023 paper, Hopster et al. introduce the notion of conceptual appropriation: the application of a contested concept to a phenomenon in order to make a claim about said phenomenon.