Par auteur > Sansò Gloria

Artificial interactive systems have become ubiquitous, ranging from virtual assistants and chatbots, capable of interpreting and responding to user queries in natural language, to automated trading systems placing trades in the stock market. These systems perform actions that are typically taken to bear the mark of the social, such as chatting and trading, prompting questions about the ontological status of these systems and their actions. Are these systems tools or social agents? Are their actions a manifestation of sociality? On the one hand, their autonomy and interactive capabilities raise doubts about the possibility of considering them as mere tools (Wein, 1992; Chopra & White, 2011; Strasser, 2020). On the other hand, they lack the cognitive abilities of an average adult human being, and this makes it difficult to see how they can be considered social agents. Indeed, in most of the literature, the consensus is that sociality requires sophisticated cognitive capabilities (see, for example, Gilbert, 1990; Searle, 1995; Bratman, 2014). In order to solve this issue, the notion of “quasi-sociality” has recently been introduced (Strasser & Schwitzgebel, 2024). According to this view, sociality is a matter of degree. Some interactions are full-blown social, whereas others are merely quasi-social because they involve partners, such as non-human animals, young children, and artificial interactive systems, that do not have advanced cognitive abilities, but still, they are somehow able to participate in social exchange. In this paper, we argue that, while the notion of quasi-sociality appears well-suited for young children and non-human animals, it may encounter limitations when applied to artificial systems. Unlike artificial systems, young children and non-human animals have the capacity for emotions. Although emotions do not entirely compensate for the lack of cognitive faculties, they play a role in genuinely connecting with others (Michael, 2010). The main aim of this paper is thus to deny the view that artificial interactive systems have quasi-social agency, and explore the alternative view that artificial interactive systems are non-social agents, meaning entities able to act with purpose, although they lack the ability to enter genuine social relations.