Send message to Author

Paper short abstract

This paper examines how “good work” is redefined in the age of AI through inter-objective relations between digital tools, platforms, and labour. Drawing on ethnography with remote IT workers in India, it shows how "LLM epistemologies" reshape "understanding" as a practice.

Paper long abstract

This paper examines how "good work" is understood in the age of artificial intelligence by asking how understanding itself — i.e., understanding as a practice—is being remade as large language models are woven into the everyday labour of remote and hybrid knowledge workers in Gurugram. I situate the argument at the intersection of the epistemologies of the digital and the anthropology of work, engaging in particular with the tradition of pragmatic genealogy (Queloz 2021). The paper's conceptual work is to distinguish three inherited ways of thinking about understanding: 1. a functionalist account of understanding as a human achievement (Hume; Weber), 2. an epistemological-realist account of understanding as more than knowledge (Grimm), and 3. an emergent techno-functionalist account of machine understanding and "LLM epistemologies". Following the literature review, I also argue that most of the literature on "understanding" remains philosophical, adjudicating what understanding is or whether machines possess it. As a result, there is little empirical account of how understanding is actually practised in everyday life. Through three thick cases, namely those of Aritra, Ritik, and Diya, et al., I show that AI-mediated understanding is reorganised as a distributed practice of prompting, bounding, and verifying. Further, I argue that this reorganisation sets two divisions moving at once: a division of labour, as the making of the thing passes to the machine and the person is left to check it; and a division of values, as worth shifts from slow, self-made understanding onto speed, fluency, and the skilled handling of the tool.