Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

The OpenModel Synthesis Framework (OMSF): A Buyer-Side Taxonomy for Evaluating AI Model Openness in Resource-Constrained and Sovereignty-Critical Deployments,

Domaine:

natural language processingdigital infrastructure

Type de record:

paper
Créateur:
Ola
Éditeur:
Zenodo
Hôte:avatar
Existing AI openness frameworks (OSAID, MOF) answer the producer’s question,  “whatmust a lab release? ”but not the buyer’s question: “can this organization legally andoperationally deploy this model in this jurisdiction?” We present the OpenModel SynthesisFramework (OMSF), a buyer-side taxonomy with four components: a six-rung OpennessLadder (L0–L5) with L1 split into sub-grades (L1a scale-capped, L1b use-restricted, L1cnon-commercial); a three-tier Source Provenance Protocol (P1–P3); a three-lens deploymentframe for private, enterprise, and non-profit buyers; and an African Edge & Sovereign Infrastructure (AESI) annotation for zero-egress feasibility under low-power hardware. We applyOMSF to N = 821 graded open-weight LLMs drawn from the top 1500 text-generationmodels on the HuggingFace Hub (snapshot 31 July 2026), after excluding quantization mirrors and library test/CI fixtures. Three findings stand out. First, repository gating isstrongly associated with license restrictiveness (χ2(1, N=821) = 86.95, p = 1.1 × 10-20,φ = 0.33): 55 of 59 gated models are L1. Second, permissively licensed models are moredownloaded than restricted-license models (Mann–Whitney U = 67,623, p = 0.002; medianL2 = 82,267 vs. median L1 = 54,852), contradicting a widely repeated procurement heuristic. Third, only 9 of 821 graded models (1.1%) declare support for any African language.Rules-variant inter-rater agreement is κ = 0.95 (almost perfect); full human IRR is planned 

Visit

doi.org

Tasks

language modeling

Licenses

Creative Commons Attribution 4.0 Internationalhttps://creativecommons.org/licenses/by/4.0/legalcode

Similaires

Agentic AI for Ethical Cybersecurity in Uganda: A Reinforcement Learning Framework for Threat Detection in Resource-Constrained EnvironmentsDefining AI Skills and AI Talent for Africa A Taxonomy, Value Chain, and Implementation FrameworkA Model for Implementing Open Badges in a Resource-Constrained EnvironmentAI Diagnostics in Resource-Constrained Settings: A Methodological Approach for MalawiAI Technology Transfer: A Five-Dimension Taxonomy and Absorptive Capacity FrameworkAntifragile Intelligence: A Triadic Framework for AI Governance, Digital Forensics, and Sovereignty in Emerging Economies

Agentic AI for Ethical Cybersecurity in Uganda: A Reinforcement Learning Framework for Threat Detection in Resource-Constrained Environments

Agentic AI for Ethical Cybersecurity in Uganda: A Reinforcement Learning Framework for

Defining AI Skills and AI Talent for Africa A Taxonomy, Value Chain, and Implementation Framework

Incorporating a Synthesis of National AI Strategies Across Africa This paper proposes that the Tale

A Model for Implementing Open Badges in a Resource-Constrained Environment

Open badges offer a unique opportunity to modularise the learning process and allow for the accredit

AI Diagnostics in Resource-Constrained Settings: A Methodological Approach for Malawi

AI diagnostics have shown promise in resource-limited settings such as those found in Malaw

AI Technology Transfer: A Five-Dimension Taxonomy and Absorptive Capacity Framework

Artificial intelligence (AI), a general-purpose technology (GPT) with infrastructure and human-capit

Antifragile Intelligence: A Triadic Framework for AI Governance, Digital Forensics, and Sovereignty in Emerging Economies

In an era defined by extreme Volatility, Uncertainty, Complexity, and Ambiguity (VUCA), artificial i