Luigia Tricase and Alice Mattoni - Contesting AI Futures: Social Movements and Solarpunk AI [Hype Studies Conference 2025 presentation]

Governments around the world are devising national AI strategies that regulate how, in the next few years, Artificial Intelligence and its technological derivatives will be implemented in different fields of public life, such as healthcare, justice, border surveillance and education. These strategies, as well as the dominant narratives about AI that they upheld, often come with promises of AI’s potential for improving democracies, one that depicts these emerging technologies as a revolution, a discontinuity from the past. National strategies promising an AI revolution are, however, often uncritically reproducing the industry utopias, discursive closure and technological sublime of narratives elaborated by the tech firms behind AI development, a mostly proprietary, venture capital-backed, black-box field of innovation.

If promises of democratic improvement are prominent among dominant narratives, they are problematized in the contentious political arena. Here, civil society actors are often developing critical accounts of the way emerging AI technologies are already employed- or plan to be- by institutions, as well as of their overall revolutionary value for societies. Among them, social movement organizations and their networks are primary developers of alternative, contentious and re-politicized visions of AI. Through their practices, these actors 1. delineate AI as an object of contention, 2. Develop alternative uses of AI as a tool for contention. Nonetheless, despite research agrees on their productive, future-making capacity, grassroots practices are currently understudied in the field of AI.

This empirical paper emerges from this literature gap, addressing it through an ethnographic approach. Starting from the case study of a social movement network active in two Italian cities throughout 2023 and 2024, we conducted participant observation and interviewed eight key actors of the network. By observing and inquiring about AI practices, our aim was to discover how- and if- AI permeates activist spaces, how it does so, and what alternative meanings about AI are being constructed by activists through their AI practices. Conclusively, we evaluated the potential of these practices to contest dominant promises and visions of AI.

While individual AI practices have been discussed with interviewees, with most of them employing Generative AI to code, write and generate images but with careful evaluation of downsides, the focus of participant observation has been on collective AI practices. Two types of collective AI practices are identified: a) present-oriented practices, and b) future-oriented practices. Through a) present- oriented practices, activists in the network elaborate alternative meanings of existing AI and its current value in societies: they discuss existing AI Errors in fields of labor, information, human rights and environmental impact; addressing these errors, they ground in present fact and re-politicize AI, rejecting dominant revolutionary narratives. Through b) future-oriented practices, activists foster imagination moments through which alternative AI futures are discussed: they experiment, through workshops and labs, desirable AIs, first by imagining and then by building and platforming prefigurative AI projects inspired by Solarpunk principles. While they remain situational and are far from implementing long-term alternative AI futures, we highlight the political potential of these practices to question dominant AI promises and utopias.