Resource-Conscious Neural Architecture (RCNA): A Decolonial AI Framework
Year: 2025
Type: article
Abstract: This paper introduces the Resource-Conscious Neural Architecture (RCNA), a decolonial, pluriversal framework that redefines artificial intelligence as "cognitive survival": the ability to achieve contextually adequate outcomes within finite ecological, computational, and epistemic constraints. RCNA is not a model, but a philosophical foundation grounded in decolonial theory, epistemologies of the South, and critiques of AI coloniality. It proposes a radical shift from intelligence as domination to intelligence as negotiated coexistence. Ethics is embedded as an infrastructural property—through mechanisms like structured forgetting, deliberative resource planning, and pluralistic governance. The Resource-Conscious Hierarchical Neural Network (RCHNN) is presented as one concrete implementation of RCNA, but the framework is intentionally infrastructure-agnostic. Alternative realizations—symbolic, neuromorphic, Indigenous-led, or community-governed—are explicitly welcomed. This is the second paper in a four-part series on decolonial AI. It follows the state-of-the-art review (DOI: 10.5281/zenodo.xxxxxx) and precedes technical and mathematical blueprints. **License**: CC BY-SA 4.0 — You may share and adapt this work, even commercially, as long as you give credit and distribute derivatives under the same license.
Author: Taleb, A
Language: English
Subfield: Safety Research
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7104024475
FWCI:
Cites:
Cited by:
Related to: