Dr Phaedra Mohammed reimagines AI, language, and literacy in the Caribbean
At the intersection of artificial intelligence (AI), education, and sociolinguistics, Dr Phaedra Mohammed is crafting technology that speaks to—and in—the voices of our region. A Computer Science lecturer in the Department of Computing and Information Technology (DCIT) at UWI St Augustine, Dr Mohammed’s work fuses advanced computer science techniques with deep cultural modelling, resulting in research with profound potential for social change.
Dr Mohammed’s journey into the world of computer science wasn’t always clear cut.
“I initially wanted to do medicine,” she recalls.
A lab experience with tissue samples changed that. “One of my cousins introduced me to computer science. I thought, ‘Hey, this is something I can probably try out’.”
It wasn’t long before she found her academic home. Her early academic path was shaped entirely at UWI St Augustine, where she completed both undergraduate and postgraduate degrees.
The real turning point came when Dr Mohammed found her passion in applied artificial intelligence—specifically, in education. Her research delved into cultural modelling and how technology could be designed to reflect and leverage the lived experiences and dialects of Caribbean users.
“I’m primarily interested in applied research that can benefit people in our society by transforming the knowledge that we discover in academia into practical tools,” she explains.
Her PhD work on culturally-aware intelligent learning environments explored whether it is possible to represent and use tangible aspects of culture in computer models that change as they learn about a student’s context. Her research showed that the use of local context made programming tasks feel more accessible to students—particularly those who felt less confident at the start.
Those early foundations in applied AI research laid the groundwork for her current flagship project: The Bright Start Project, a regional initiative aimed at developing an intelligent literacy tutor designed specifically for Caribbean children. The project, with parallel arms in Trinidad and Jamaica (at UWI Mona), is a response to the widening reading gap experienced by children in primary schools.
“A common benchmark is that children should be reading fluently by the age of 10. Recent World Bank reports show that Trinidad and Jamaica are well below global reading thresholds,” she notes.
Bright Start tackles this issue using AI, speech and language technologies, and instructional modelling.
“It’s designed for the Caribbean context. The tutor listens to a student’s oral reading, analyses and grades the reading performance for accuracy and fluency, and then gives detailed feedback of reading errors (if any), with the option for the student to replay their reading,” Dr Mohammed explains.
Crucially, the system is designed specifically to recognise regional accents and local speech patterns.
“Right now, educational software coming out of the Global North, doesn’t cater for our unique accents. So, if a student says ‘de cat is red’, the pronunciation of ‘the’ is flagged as incorrect. But ‘de’ is a Trinidad English Creole pronunciation of ‘the’. It’s not incorrect as it’s acceptable in informal learning settings.”
The tutor, therefore, doesn’t shame but rather guides with responses such as “try saying it this way”— an approach that affirms cultural identity while gently steering students toward formal literacy expectations.
Backed by impressive grants—including Jamaica’s Universal Service Fund and the US-based Future Forum Learning Engineering grant—the team has already rolled out AI-driven prototypes of their literacy tools.
“We tested a spelling tutor with several students and saw increases in their speed and correctness of spelling. We were even able to detect creolised spelling—words spelled how they sound,” Dr Mohammed says.
The team’s efforts also involved collecting speech samples to train the Bright Start reading tutor to better recognise regional voices. “We collected over 400 voice samples from children reading in Trinidad—from Moruga to Mayaro to Arima, St Ann’s and Chaguanas.”
To take it a step further, the team is experimenting with voices that sound authentically Trinidadian in place of the off-the-shelf options widely available.
A core vision of the project is deep personalisation: “If you’re from a coastal village or a refinery town, the tutor may use that context to recommend reading passages about sea creatures or mining activities. A student would know particular domain-specific words and might better contextualise their learning through cultural anchors.”
Dr Mohammed’s collaborator, Dr Andre Coy, Associate Dean and Senior Lecturer at UWI’s Mona Campus, initiated the literacy-focussed concept with speech technologies, while she brought in the cultural modelling and intelligent tutoring systems expertise. Together with a diverse team at The UWI, Dr Paulson Skerrit, Dr Yewande Lewis-Fokum and Mr Asad Mohammed, they’re pushing the boundaries of educational AI in the Caribbean.
Thanks to the explosive growth in computing power, the speed of tech innovation is staggering. Yet, with all the talk of AI giants, Dr Mohammed’s work stands out for its empathy.
“Many of the mainstream tools flag our speech as wrong. But we’re not wrong—we’re different—and that is worth exploring and celebrating,” she says.
As the Bright Start Project moves toward a public prototype launch later this year, it represents more than just a tech milestone. It embodies a future where Caribbean children learn in their own voice, at their own pace, and in their own context.
For Dr Mohammed, this is not just computer science. It’s personal, it’s transformational and it’s Caribbean-centric. “It’s meant to be a learning companion,” she says, “where a child can read, and the tutor will guide them - speaking in a voice they understand.”
A voice, that quite possibly for the first time, will sound a lot like home.
