Why AI Needs to Understand Emirati Arabic, and Who Is Building It
Zeta42 · 18 September 2026
MBZUAI's Dr Hanan Aldarmaki researches how AI handles Arabic and its dialects, as Ramsa Lab documents spoken Emirati Arabic to preserve heritage and build datasets.
A local question at the heart of UAE AI
The UAE's AI ambitions reach beyond algorithms and computing infrastructure to the people developing the technology. A new profile in Fast Company Middle East looks at one of them, and at a question that turns out to be deeply local: can AI accurately understand the people it is built to serve?
That question drives the work of Dr Hanan Aldarmaki, Assistant Professor of Natural Language Processing at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) and director of its Center for Teaching and Learning. When she joined MBZUAI in 2022, she became the university's first woman and first Emirati faculty member. Today she researches Arabic language and speech technologies, teaches and supervises AI researchers, and leads the university's teaching and learning efforts.
Why Arabic dialects are hard for AI
Aldarmaki's research covers natural language and speech processing, with particular attention to Arabic and to low-resource languages and dialects. Arabic is a complex challenge for AI because it has many regional dialects. Spoken Emirati Arabic adds another layer of nuance on top of that.
The practical consequence is easy to picture. A voice assistant, chatbot or search tool is only as useful as its grasp of how its users actually speak. For anyone building generative AI products for Gulf users, dialect is a product decision. A system that misses the way people talk at home will serve some users far better than others.
What Ramsa Lab is documenting
Capturing those differences is the focus of work being carried out through Ramsa Lab, a national initiative dedicated to documenting spoken Emirati dialects. According to Fast Company Middle East, the project goes further than teaching AI to recognise different ways of speaking.
Much of the UAE's intangible heritage has traditionally been passed down orally. Recording and documenting spoken language therefore does two jobs. It helps preserve cultural knowledge, and it creates the datasets and resources that allow AI systems to represent Emirati language and culture more faithfully. The same recordings serve the archive and the model.
Why identity has become a technical issue
The overlap between technology and identity matters more as AI systems shape search, translation, education and digital assistants. These are the tools people use to find information and to learn. If they handle Emirati Arabic poorly, Emirati speakers get a weaker version of each of them. If they are built on well-documented local speech, the dialect becomes something the software can work with instead of around.
From AI's classroom to its research labs
Aldarmaki's career reflects a wider shift. Emirati women are increasingly moving into scientific research and advanced technology, rather than taking part in those sectors only as users or professionals. According to Fast Company Middle East, female students account for 54% of the Emirati student population in MBZUAI's 2026–27 cohort.
Her own route ran from the UAE to postgraduate study in the UK and the US. She holds an MPhil from the University of Cambridge and a PhD in Computer Science from George Washington University, and she brought that expertise home. Now she is helping build the pipeline for the researchers who will follow. In her view, that ecosystem is the product of sustained investment in education, research and women's participation in the workforce.
Her advice to girls: build expertise
Aldarmaki's message to girls considering careers in AI is direct: build expertise. Understanding the language, concepts and technical foundations of a field gives women more than an entry point into technology. It gives them a voice in deciding where that technology goes next, which matters as the UAE positions AI as a pillar of its future economy.
That advice travels well beyond one university. Zeta42 teaches AI to children, teens and adults on the same principle: using the tools is where learning starts, and understanding how they work is what lets someone question them and build with them. For parents of a teenager who is curious about AI, this story is also a useful reminder that some of the most interesting work in the field is about language and culture as much as code.
A tool, not the destination
Aldarmaki also sees a limit to the technology-first narrative. AI, she argues, is ultimately a tool and not the destination. The challenge for the next generation of researchers is to make sure the systems they build contribute to human prosperity and a sustainable future, and greater capability alone does not guarantee that.
For Emirati women entering AI, that means a role in deciding what the technology's future looks like, as well as in adopting it. An AI that can understand spoken Emirati Arabic is one concrete place where those decisions are being made now.
Source: fastcompanyme.com
