Negotiating alternative AI futures: a critical engagement with European civil society organizations, tech entrepreneurs and journalists
Type
conference paper
Date Issued
2024-09-26
Abstract
With the popular diffusion of artificial intelligence (AI) systems in our society pushed by the rise of generative AI systems such as ChatGPT and Midjourney, the critical debate around AI has exited expert and technical circles. The existing debate on AI is often hyped and polarized and characterized by either catastrophic or utopian points of view, which are directly overshadowing the discussion of the social disorder emerging from the concrete harms for which AI systems are already responsible (Barassi et al., 2022). Currently, most of the debate around AI is thus driven by the discussion of "existential threats" which are mostly speculative in nature, and not by a coming to terms with already documented and visible issues, such as discrimination, biases, or data (in)justice perpetuated or amplified by prone-to-errors algorithmic systems (Ryan-Mosley, 2023). With the aim of writing an "history of the present" (Aradau and Blanke, 2021) of the social construction of AI in this crucial moment, in this paper, we thus investigated how three crucial social groups are coming to terms with the fallacies of AI systems and how AI "errors" could also serve as starting point for negotiating our possible futures with AI systems (Barassi, forthcoming). Through 105 in-depth interviews with civil society representatives, tech entrepreneurs and journalists based in Europe, this paper aims at 1) mapping the concerns of three of the most engaged groups involved in the debates around AI. Additionally, through the analysis of their views and discussions of the various "errors" of AI, the paper also wants to 2) contribute to the analysis of how the critical responses to AI harms are being currently expressed and negotiated by different critical societal actors. At the same time, the paper also discusses 3) alternative AI futures, where expert views and takes on the technology are addressed towards the building of alternative future scenarios, where both dystopian and techno-fetishist views get challenged in their most evident weaknesses and rhetorical shortcomings. Overall, the paper reacts to hegemonic and normative takes on AI and its futures, looking beyond the utopia/dystopia juxtaposition, which is usually presented as ineluctable by different stakeholders (Bory, 2019). The results presented in this paper emerge from "The Human Error Project", a multidisciplinary research project launched in 2020 with the goal of studying how different social actors are making sense of AI systems through an analysis of where the debates around the errors of these systems are driving our collective and societal construction of AI. Blending anthropology and media studies approaches and methodology, "The Human Error Project" and this paper position themselves among the growing literature and research field of critical data and AI studies.
Language
English
Keywords
AI error
artificial intelligence
anthropology
media studies
critical data and AI studies
Event Title
ECREA (European Communication Research and Education Association) 2024 Conference - Communication & social (dis)order
Event Location
University of Ljubljana, Slovenia
Event Date
24-27 September 2024
Subject(s)