Anthropomorphization and Materialization: Good Prac-tices within an Applied Ethics Framework for an AI Risk Taxonomy

Moving from the risk of individuals in over-simplifying the representation of reality due to internal biases and external process of disinformation (i.e. “epis-temic vulnerability), this chapter will focus on two main ethical processes to turn the representation of AI from neutral to a end-oriented one . Grounded on a philosophical framework which brings together epistemology and applied ethics, (a) anthropomorphizing and (b) materialization of AI will be assumed as good practices to face socio-cultural and enviromental threats raised by Ar-tificial Intelligence. Although these two concepts show a long and usually neg-ative cultural heritage, we will try to rethink them in a positive perspective against mainstream narratives about AI. Indeed, anthropomorphization can support in representing rewards and algorithms-based platforms as produced by companies aimed by specific commercial ends with hidden social costs (e.g. micro-labour, new form of unconscious slavery). Furthermore, materialization can turn our biased perspective of “immaterial” digital environment as strong-ly related to abuse of natural resources with hidden environmental costs. In conclusion, this chapter wants to introduce a revision-bias ethical strategy to avoid false ethical problems raised by AI and focus on the present and real ones.