For decades, diagnosing and treating mental health challenges in Africa has been hindered by limited data, cultural mismatches in diagnostic tools, and a severe shortage of trained professionals.
But with advancements in technology, Shamiri Health has developed an AI platform designed to make mental health diagnosis and treatment more accurate, affordable, and accessible for millions of young people.
The technology, developed in Kenya and trained on African data, represents a new chapter in global mental health innovation—one driven from within the continent.
Tom Osborn, the chief executive and founder of Shamiri, said that historically, mental health diagnosis relied on Western-designed screening tools and human clinicians, both of which are scarce and costly in Africa.
He noted that even when screening tools are used, cultural and linguistic nuances often distort results, making it difficult to understand the true scale or nature of mental health challenges.
In Kenya, fewer than 5,000 psychiatrists serve a population of over 50 million people, leaving most young people without access to formal care.
To change that, the Shamiri Institute has combined artificial intelligence, data analytics, and community delivery to bridge the gap.
A standout feature of ShamiriAI is its Automatic Speech Recognition system, trained on English, Swahili, and Sheng—the languages actually spoken by youth in therapy settings.
This allows the system to interpret tone, emotion, and nuance that Western AI models often miss, making it one of the first culturally attuned mental health AI tools in the Global South.
By embedding this technology into its existing peer-led school model, where 18- to 22-year-old youth deliver evidence-based interventions, Shamiri has achieved remarkable results: up to an 80 percent reduction in anxiety and depression symptoms, sustained over three years.
“AI doesn’t replace human empathy but instead enhances it. By empowering peer providers with insights they wouldn’t otherwise have, we make care smarter and more human at the same time,” Osborn said.
This allows interventions to be both more personalized and culturally grounded—a leap forward in diagnostic precision.
“AI allows us to bridge gaps that human systems alone cannot close. It makes care smarter, faster, and more contextually relevant,” said Osborn.
He added, “For too long, Africa’s mental health data and experiences have been invisible. Now we can use technology to understand, predict, and respond to youth mental health in real time.”
According to findings from a survey conducted between 2021 and 2023, researchers found that 30 per cent of adolescents met the threshold for depression and 25 per cent for anxiety—alarming rates that fluctuate with social and academic pressures.
Faith Kamau, Research Officer at Shamiri Institute, notes that the numbers peaked at 42 per cent for depression and 38 per cent for anxiety after the COVID-19 pandemic, before rebounding in 2023.
She said this indicates that mental health symptoms are responsive to broader social conditions.
“Between 2021 and 2023, Shamiri surveyed 7,865 secondary school students across 27 schools in four counties. Mental health cases increased during the COVID-19 period compared to before and after,” she said.
Kamau highlighted that final-year students and girls in single-sex schools were at the highest risk, with exam stress and social isolation playing major roles.
“Academic performance, social support, and a sense of personal control were found to be powerful protective factors among those who took part in the survey,” she said.
Kamau said that researchers also used network analysis, a data-driven approach akin to AI mapping, to identify the most influential symptoms within mental health networks.
“Worry emerged as the most central symptom. Hopelessness and nervousness acted as bridges between depression and anxiety. This means that helping students manage worry or address hopelessness could reduce the entire web of symptoms,” Kamau explained.
These findings not only quantify the crisis but also point directly to where AI-driven interventions can have the greatest impact.
ShamiriAI’s machine learning models are designed to detect such symptom patterns early and recommend context-specific support—something human therapists alone cannot scale.
“When I look at this data, I don’t just see numbers. I see students balancing heavy expectations with quiet courage. Many are struggling silently, but they are also incredibly resilient. Our job, as researchers, educators, and leaders, is to make sure that resilience is met with understanding, support, and opportunity,” she noted.
International leaders have hailed the initiative as a milestone for global health equity.
H.E. Arnaud Suquet, French Ambassador to Kenya and Somalia, praised Shamiri for showing how AI can power affordable, scalable healthcare models.
“Kenya is one of Africa’s leading tech hubs with the talent and green energy to drive sustainable AI systems. Through innovations like ShamiriAI, Shamiri is showing that technology can truly reach young people where they are, and at a cost ten times lower than traditional therapy,” said the Ambassador.
By 2032, Shamiri Institute aims to reach 10 million youth across Africa through its AI-enhanced programs.
“Africa doesn’t have to wait for imported innovation. We are training AI on our languages, building on our data, and creating technology that understands our people. That’s the real power of African innovation,” Osborn emphasised.
