While attending a regional training on ethical and responsible AI in Sub-Saharan Africa, I was drawn to reflect deeply on this question. The training aimed to enhance understanding of Artificial Intelligence (AI) through the lenses of ethics, responsibility, and real-world applications, particularly in health.
It emphasised that ethical AI ensures systems are designed and deployed in accordance with social values such as fairness, justice, transparency, and non-discrimination values that promote equitable and responsible outcomes. However, it also became clear that AI systems should not merely claim to represent these principles; they must actively demonstrate them in how they operate and impact society.
This realisation prompted a deeper reflection: has Uganda, as a nation, defined the ethical and moral foundations that should guide AI development and use within its unique social and cultural context? Ethics are essentially the moral principles that guide human behaviour and are often reflected in the laws, norms, and expectations of society. If Uganda is to benefit fully from AI, we must define a local ethical framework rooted in our cultural values of fairness, community, inclusivity, and human dignity.
Without this, AI applications risk importing biases from foreign systems, leading to inequitable or even harmful outcomes. For example, an AI healthcare model trained on data from Western populations may misdiagnose diseases or recommend ineffective treatments for Ugandan patients due to genetic, environmental, or infrastructural differences.
Ethical AI, therefore, must not be imported, it must be contextual, practical, and human-centered. One of the most powerful lessons from the training was the principle of protecting human autonomy. AI should never undermine human control or replace human judgment. Instead, it should supplement human intelligence and decision-making capacity.
I often relate this to the introduction of the Video Assistant Referee (VAR) system in football. VAR uses AI-assisted technology to help referees make better decisions, but it does not replace them. The final judgment always rests with the human referee. Similarly, in every domain where AI operates whether in education, governance, or health, the human element must remain central. Humans must validate AI outputs, correct errors, and ensure technology operates within ethical and moral boundaries.
Without human oversight, AI risks becoming an unmonitored system that makes decisions without empathy or accountability. This leads to another key dimension: AI must prioritise human well-being, safety, and the public interest. The purpose of AI should never be to cause harm or to devalue human dignity. Instead, AI should enhance lives by improving efficiency, accuracy, and accessibility of services while protecting users from discrimination, stigmatisation, and data exploitation.
The true measure of progress lies not in how intelligent machines become, but in how responsibly and equitably they serve the people. Equally important is the concept of responsibility and accountability. AI systems do not exist in isolation, they are created, trained, and managed by humans. This means accountability should be clearly defined at every stage of their lifecycle.
Developers, policymakers, and users must all understand what the system can and cannot do, under what conditions it operates, and what mechanisms exist to detect and correct its errors. When AI causes harm, there must be clarity on who bears responsibility, the developer, the institution, or the user.
Without such mechanisms, AI can become a ‘black box,’ making life-altering decisions that cannot be explained or challenged. Transparency is, therefore, crucial. Developers should disclose data sources, model limitations, and decision-making criteria. Regular audits and human supervision should be embedded into AI systems to ensure accountability is not an afterthought but a core design feature. While AI’s potential is immense, it is vital to recognise its limits. AI is not a magic solution that can automatically fix all societal problems.
It functions within the boundaries of its design, training data, and human instruction. Over-reliance on AI without critical evaluation can lead to unintended consequences. Therefore, AI should be viewed as a partner in human progress, a tool that extends human capacity rather than replaces it. Uganda must adopt a balanced approach: embracing the opportunities that AI presents while critically managing its risks.
This involves developing national policies that regulate AI’s use, encourage ethical innovation, and promote inclusivity. It also calls for investment in local research and training to ensure Ugandan experts can design systems that reflect our social realities and needs. Collaboration between government, academia, and the private sector is essential to build an AI ecosystem that is both innovative and responsible.