Metabolic Infrastructures

 

Artificial intelligence is not an autonomous intelligence hovering above material reality; it is a political metabolism that converts energy, land, labour, cultural expression and institutional trust into computational capacity. Data centres expose this dependence by placing model expansion within regional struggles over electricity, emissions and public infrastructure, while data feminism reveals that datasets, classifications and labour arrangements already embody structural power. Copyright disputes likewise demonstrate that training is not a singular act but a sequence of acquisition, copying, curation, memorisation and retrieval whose legitimacy depends upon purpose, provenance and market consequence. Within Socioplastics, these conditions require the corpus to be conceived not as an immaterial archive but as a cultural commons sustained by inspectable gateways between sources, machines and publics. A meaningful node must retain durable identity, permissions, contextual density and intellectual lineage as it migrates across repositories and interfaces. Persistent identifiers, interoperable formats, redundant storage, licensing statements and concise metadata therefore constitute ethical as well as technical devices. For example, an index may enhance discoverability while suppressing alternative relations; similarly, an API may widen access yet privilege computationally equipped actors. The project’s fifteen-year installed base of posts, exhibitions, images and institutional experiments provides a decisive case study: reactivation must preserve temporal and authorial difference rather than flatten heterogeneous materials into anonymous training feedstock. CitationalCommitment consequently names an infrastructural obligation to keep every transformation—from source to concept, operator, index and circulation—attached to its origins. By joining machine legibility to human contestability, Socioplastics establishes a counter-model in which scale becomes negotiable stewardship rather than concealed dispossession.

Anto Lloveras is an architect and urban researcher working across spatial practice, epistemology, media archives and public infrastructures through LAPIEZA-LAB and Socioplastics.

Herremans, D. (2025) ‘Music Royalties: How to Treat Artists Fairly in the Age of AI Music’, WIPO Magazine, 2025 Music Special, pp. 48–53.

Klein, L.F. and D’Ignazio, C. (2024) ‘Data Feminism for AI’, manuscript submitted to ACM FAccT 2024, arXiv:2405.01286.

Marttila, S. and Botero, A. (2017) ‘Infrastructuring for Cultural Commons’, Computer Supported Cooperative Work, 26, pp. 97–133. doi:10.1007/s10606-017-9273-1.

Radcliffe, D. (2025) Journalism in the AI Era: Opportunities and Challenges in the Global South and Emerging Economies. London: Thomson Reuters Foundation.

Sonabend, R., Gruson, H., Wolansky, L., Kiragga, A. and Katz, D.S. (2024) ‘FAIR-USE4OS: Guidelines for Creating Impactful Open-Source Software’, preprint.

Springer, C. and Hasanbeigi, A. (2025) Data Centers in the AI Era: Energy and Emissions Impacts in the U.S. and Key States. San Francisco: Global Efficiency Intelligence.

Taylor, G. (2025) ‘Artificial Intelligence: The Synergy of Technology and Creativity’, WIPO Magazine, 2025 Music Special, pp. 44–47.

United States Copyright Office (2025) Copyright and Artificial Intelligence, Part 3: Generative AI Training. Pre-publication version. Washington, DC: United States Copyright Office.

Vázquez, M.L. (2025) ‘Could AI Music Be the Industry’s Next Napster Moment?’, WIPO Magazine, 2025 Music Special, pp. 34–39.