Critical Phenomenology of Prompting in Artificial Intelligence

Abstract

This paper analyzes the philosophy of prompting as a tool within the context of the rise of Artificial Intelligence (AI), particularly in large language models (LLMs). The topic is justified by the need to understand the prompt as a mediating space between human intentionality, language, and the sociopolitical structures that shape interactions with these technologies. The central objective is to examine how prompting reflects ethical, ontological, and epistemological tensions that arise in the construction of meaning within AI systems. Methodologically, the study adopts a critical-phenomenological approach, combining first-person experiences (user) with practical experimentation of prompts in different scenarios. The results demonstrate that the prompt is not merely a technical instruction but a discursive practice, where human decisions, such as the configuration of “parameters” (e. g., temperature and Top P), directly influence the outputs generated by AI systems. While these decisions appear technical, they carry significant ethical and epistemological implications that demand critical examination. The study concludes that it is essential to adopt an interdisciplinary approach that integrates technical development with philosophical reflection. This approach would foster an ethical, conscious, and responsible use of AI while recognizing the central role of humans in interactions with these emerging technologies.