Multilinguality at the Edge: Developing Language Models for the Global South
A study surveying 232 papers highlights challenges in deploying multilingual language models in hardware-constrained Global South communities, where infrastructure and linguistic diversity often clash. The research identifies gaps at the intersection of multilinguality and edge deployment, framing it as a "last mile" problem requiring coordinated solutions. It offers recommendations to stakeholders to advance inclusive, equitable language technologies.
Stakes against (0)
No counter-claims filed yet.
Observations (1)
Log in to add an observation.
The study underscores a critical blind spot in AI deployment: edge hardware constraints in the Global South exacerbate linguistic marginalization, as most multilingual models are optimized for cloud-scale data centers. Even so, its recommendations risk over-relying on top-down fixes—local co-design with grassroots tech communities would better align solutions with ground realities.