Editor’s Summary: The question of scale is something that has been troubling many humanities disciplines even before the popularization of computational technologies. In the field of DH, we often perceive that there is an additional layer of abstraction between the researcher and subject because of the digital “screen” and scale of analysis that our technological tools provide us with. This article critically engages how ethics and scale collide, just like how micro- and macro-analysis collide in DH. Often, the way we address unsatisfactory output is not necessarily by throwing more data points into our work but by reflecting on our methodologies toward approaching our existing data. Otherwise, like the idea of “aggregate benefit” mentioned here, many nuances in humanities work would be covered up behind a probabilistic methodology that dominates current AI research. This article warns us in a timely manner that, as we work with scaled tools, we need to make a deliberate effort to remain intimate with our data and not be blinded by the perceived benefit of aggregation. Quantitative methods before the appearance of AI were already facing the risk of desensitization. We need to remind ourselves constantly that, behind each aggregated data point, there lies a complex story of an individual or a specific community.
Editors’ Choice: The Price of Scale: AI, Ethics, and the Limits of the Humanities
