Google.org backs MBZUAI research on underrepresented languages
Research targets MENA’s diverse linguistic landscape
#UAE #LLMs – Google.org has awarded $1 million funding to Abu Dhabi-based Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) Vice Provost of Faculty Excellence and Advancement and Professor of Natural Language Processing Dr. Thamar Solorio to lead a research initiative addressing the data divide preventing AI from serving speakers of underrepresented languages with accuracy comparable to English. The initiative focuses on developing resource-lean AI training frameworks requiring less manually annotated data and lower computational power.
SO WHAT? – The MBZUAI research addresses a fundamental barrier in local AI development. Language data used to train models remains heavily skewed toward data-rich Western languages, causing language models to lose cultural nuance and technical precision when applied to MENA’s linguistic complexity. Meanwhile, it is often costly and time-consuming for developers to overcome this barrier. MBZUAI aims to democratise innovation by enabling local institutions, start-ups and researchers to develop sophisticated AI tools without requiring substantial capital or infrastructure. The research will establish a native framework grounded in sociocultural and linguistic realities of the Middle East and North Africa (MENA) region, with a focus on Arabic.
Here are some key points about the research and Google.org grant:
Google.org has awarded $1 million funding to a professor at Abu Dhabi-based Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) to develop inclusive, high-performance AI for the region’s diverse linguistic landscape.
Vice Provost of Faculty Excellence and Advancement and Professor of Natural Language Processing Dr. Thamar Solorio will lead a research initiative addressing the data divide preventing AI from serving speakers of underrepresented languages with accuracy comparable to English.
Hundreds of millions of people across the Middle East and North Africa speak Arabic dialects and minority languages that are poorly supported by today’s language and speech technologies. The MBZUAI project aims to address the gap by rethinking how AI systems learn language, moving beyond approaches designed for English and other high-resource languages.
The funding supports development of resource-lean AI as global models become increasingly massive and expensive to maintain, with new training frameworks designed to democratise innovation by lowering barriers to entry for AI development across the MENA region.
Beyond technical frameworks, the funding will enable talent training, supporting a new generation of postdoctoral and early-career researchers with sustained leadership necessary to build a field that has historically been underrepresented in AI research.
Professor Solorio’s team is establishing research framework grounded specifically in sociocultural and linguistic realities of the MENA region rather than adapting high-resource models built primarily for Western languages and contexts.
For example, due to complex linguistic structures and diverse dialects, a single sentiment can be expressed in multitude of ways. Building a framework to account for such linguistics could have an impact on education, cultural preservation and digital communication applications.
The research could empower communities across MENA with speech and language technologies tailored to specific needs, ensuring no community is left behind in AI evolution.
MBZUAI is the first graduate-level university dedicated entirely to advancement of science through artificial intelligence.
The research grant aligns with Google.org’s AI Opportunity Initiative commitment to providing access to innovative AI technology in Arabic, its dialects and other languages spoken in the region through collaboration with institutions deeply rooted in advancing AI research.
ZOOM OUT – The MBZUAI funding forms part of Google.org's AI Opportunity Initiative for MENA, launched in 2024 with a planned $15 million contribution to AI-related projects between 2024 and end of 2027. Since 2019, Google.org has provided more than $25 million to organisations across MENA, including over $17 million in cash funding, $8 million in in-kind giving and 7,000 pro-bono hours of Google employee time. Previous AI Opportunity Initiative investments include $1 million to Micromentor scaling digital mentorship for 21,000 entrepreneurs contributing to 3,800 job creations, $1 million to International Water Management Institute developing AI-powered treated wastewater platform for agriculture, and $1.5 million to INJAZ Al-Arab empowering 160,000 youth with business education and online safety skills across the Middle East and North Africa.
[Written and edited with the assistance of AI]