Responsible AI : The Intertwine Between Latour's Actor-Network Theory, Aristotelian Conditions of Responsibility, and Dooyeweerd's Modal Aspects

Abstract

Currently, the discourse surrounding Artificial Intelligence often revolves around two polarising viewpoints. Exaggerated beliefs in AI’s intelligence have led to the idea of anthropomorphised AI as a new digital species. Nevertheless, there are concerns regarding the potential of AI to pilfer personal data, stifle creativity, eradicate livelihoods, and maybe lead to the downfall of human society. In order to avoid both approaches, which are either fearmongering or fawning towards AI, I propose an approach that emphasises the crucial involvement of humans in designing, training, and implementing responsible AI. My proposal is built on the integration between Latour’s Actor-Network Theory (ANT), Dooyeweerd’s modal aspects, and Aristotelian principles of responsibility. The first step is to identify the actors (human and non-human) and their modal aspects. The second step is to map the interactions between actors within the network. The last step will apply the Aristotelian principles of responsibility (knowledge, control, and freedom) to each actor. By combining these three perspectives, we can establish a framework for responsible AI based on Aristotelian principles, taking into account the constraints of each actor’s modalities and their collaborations within the network of systems.

Watch on YouTube:

https://youtu.be/b6yyQU7lx9Y?si=2w10n5ORh40hIqqC