AI’s Quiet Rise in Uganda: Challenges to Scaling Success
Artificial Intelligence (AI) has been hailed as a revolutionary force akin to electricity or the steam
engine. It is evolving in leaps and bounds worldwide, it is being embedded within various
sectors on a large scale to great benefit. Uganda, like many other countries, could benefit greatly
by leveraging AI. To unlock its huge potential, the hurdles to scaling up must be overcome.
Below are some of the challenges that Uganda faces.
1. Data insufficiency
AI feeds on data. It has been compared to a car. Like a car that relies
on fuel, AI relies on data such as numbers, pictures, words, videos ,etc. This information
enables AI to learn patterns. From the patterns it has learned, it is able to process new
data and produce new results. A fictional scenario of this would be an AI traffic assistant
used by Uber drivers in the city. Every day it is fed information from traffic camera videos,
road speeds, and weather reports; because of this, it learns that certain roads clog up
with traffic whenever it rains. Therefore, when a new rainy day comes, it processes that
information and advises drivers to take different routes. Unfortunately, Uganda lacks a
robust data collecting culture. Even when data is collected, it is in paper form and it is
often poorly stored and easily lost. Furthermore, given AI’s data reliant nature, a reliance
on datasets from outside Africa which lack the patterns and contexts unique to Uganda
means that those models will not be accurate because they reflect foreign conditions.
2. Technological limitation
Going back to the analogy of the car. A car requires an engine
because an engine is where fuel is burnt to create the power that moves the car. AI’s
engine is computing power. There is a need for computers that can process tonnes of
data in an instant. Added to that is storage capacity for all the data that the AI needs to
work with. The greater the computing power and storage capacity, the more powerful
the AI. In 2023, the UAE-based G42 unveiled Condor Galaxy, a global network of nine
interconnected AI supercomputers with a planned capacity of 36 exaFLOPs. Returning to
the car analogy, each Condor Galaxy supercomputer is like a powerful car. Alone, each
car is powerful, but together, they are even more powerful and capable of achieving far
more. Closely connected to computing power is the need for a steady supply of
electricity, which is still a pipe dream in many parts of Uganda. Without investing in
supporting infrastructure, realising the benefits of AI on a large scale will remain wishful
thinking.
3. Costs, capacity and funding barriers
Building an AI friendly environment requires robust
data collection, high speed internet and powerful computing systems. There is also a
need for specialised training in order to bridge the skill gap required to boost the uptake
of AI in the country. These factors require enormous investment, which may not be a
priority for the government.
4. Resistance
Many Ugandans are, by nature, deeply suspicious, and old-fashioned. When
they hear about AI, many of them are not willing to investigate the subject. They are also
ignorant of the uses to which AI is being harnessed globally. For instance, in China, the
Squirrel AI personal tutoring app assesses where a child is struggling and tailors lessons
to their pace. It begins with a test to assess a student’s strengths, weaknesses, and
learning speed in a given subject. Then, it simplifies that subject into small components
called “knowledge points.” These points are linked in a "knowledge graph" that indicates
how these concepts are related and support each other. The system then generates
lectures through brief videos, examples, and practice problems while adjusting its
difficulty as the student progresses. If students quickly exhibit a better grasp of a topic,
the system advances them; if they struggle, it slows down and provides more support.
Another example of AI’s transformative uses was when Google’s DeepMind collaborated
with Moorfields Eye Hospital to assess eye scans with the former’s AI system. The
system was able to correctly recommend how patients should be referred for treatment
for over 50 sight-threatening eye ailments as accurately as the world-leading expert
doctors. Closer to home, in Kenya, AI was successfully integrated into agriculture
through the Plant Village Nuru Application. The application utilised digital assistance
technology to aid farmers in diagnosing crop diseases in the field. TensorFlow image
detection technology was used to identify objects and patterns on plant leaves
indicating disease outbreaks. The application did this by training on images of both
healthy and diseased crop leaves. The examples above show AI’s malleability and
potential to positively impact various sectors.
5. Policy framework absence
Uganda is yet to craft a national policy addressing AI.
Mauritius, Egypt, Zambia, Tunisia, and Botswana have all taken the bold steps of creating
national AI policies while Uganda is yet to follow suit. A national policy would prioritise
AI development specific to Uganda’s unique conditions; drive economic growth by
boosting productivity and facilitating innovation. It would also help to reduce costs,
provide clear rules for how AI will be governed, grow local AI talent, and keep Uganda
competitive on the world stage.
AI is no longer the future. It is here now and it is here to stay. Harnessing its potential on a large
scale will only happen when these significant barriers are dismantled.