Moffat Mungazi
Dr Evans Sagomba
Everything AI
THE hum of technological advancement resonates globally, and Zimbabwe, with the recently concluded Zimbabwe National Artificial Intelligence (AI) Strategy, 2026-2030, there must be talk of transitioning from the strategy to policy.
As we chart this course, a critical question looms: How do we ensure that the AI systems we adopt serve the interests of all Zimbabweans, mitigating potential harms and fostering equitable progress?
The answer, I believe, lies in making a thorough understanding of “high-risk AI” the cornerstone of the required National AI Policy.
Why this emphasis on risk?
Simply put, ignoring the potential downsides of AI is akin to building a house on a shaky foundation.
Let us consider for a moment the concept of “high-risk AI.” It is not a straightforward, universally agreed-upon definition.
Instead, it requires a nuanced assessment of the potential impact of AI systems on individuals, groups, and society as a whole.
What types of AI models and systems should be under stricter legal scrutiny and oversight?
How do we assess if an AI system or model poses a risk to Zimbabweans?
Imagine, for instance, the deployment of facial recognition technology by law enforcement. While it may offer enhanced security capabilities, it also carries the risk of misidentification, bias against certain ethnic groups, and erosion of privacy. Similarly, AI-powered loan applications could perpetuate existing inequalities if algorithms are trained on biased data, unfairly denying access to credit for marginalised communities. The potential for misuse and unintended consequences is vast.
Crucially, we must recognise that “Artificial Intelligence” is an umbrella term encompassing a diverse range of technologies, each with its own unique risk profile.
A general-purpose chatbot, a self-driving car, a voice assistant, a home humanoid robot, a security camera, a toy, and a social media recommender system might all fall into the AI category. But these technologies are different, affect different people, are used in different contexts, and clearly present different risk profiles. Even though they might fit the same “AI” category, they probably should be regulated differently.
Let us consider a chatbot designed to provide medical advice. While seemingly benevolent, inaccurate or biased information could have serious health consequences.
Contrast this with an AI system used to optimise agricultural yields.
While the risk of direct harm to individuals might be lower, the potential for environmental damage or economic disruption still warrants careful consideration.
This is one of the challenges of AI regulation, as products that have almost nothing in common will be under the same legal framework and might have to follow the same rules simply because they have an AI component or they fit the law’s definition of AI.
Therefore, a blanket approach to AI regulation is not only impractical but also potentially counterproductive. We need a framework that differentiates between high-risk and low-risk AI applications, tailoring regulatory oversight accordingly.
What constitutes high-risk AI?
Defining “high-risk AI” requires a deep understanding of the specific challenges and vulnerabilities facing Zimbabwe. This cannot be a mere cut-and-paste exercise from international frameworks like the EU AI Act, which, while commendable, may not fully capture the nuances of our local context. Existing attempts to classify and regulate AI systems as risky or not-risky have been incomplete, rigid, and have overlooked essential aspects of the interaction between humans and machines. Even the EU AI Act, often seen as ‘strict’ or a global reference in AI regulation, misses critical AI-related risks that have become evident over the past three years and have already led to multiple cases of harm, including death.
Instead, we must engage in a robust, participatory process involving policymakers, legal experts, technologists, civil society organisations, and, most importantly, the communities that will be directly affected by AI technologies. This dialogue should address the following key questions:
What are the potential harms that AI systems could inflict on individuals and communities in Zimbabwe? This includes not only physical harm but also economic, social, and psychological harm.
Which sectors are most vulnerable to AI-related risks? Consider areas such as healthcare, education, finance, law enforcement, and employment.
How can we ensure that AI systems are used in a way that promotes inclusivity and equity? This requires addressing potential biases in data and algorithms, as well as ensuring that marginalised communities have access to the benefits of AI.
What mechanisms are needed to hold developers and deployers of AI systems accountable for any harm caused? This includes establishing clear lines of responsibility and providing effective remedies for victims.
Based on these discussions, we can then develop a comprehensive definition of “high-risk AI” that is tailored to the Zimbabwean context. This definition should be dynamic, evolving as our understanding of AI and its potential impacts deepens.
Building a robust AI Policy framework
A robust AI policy framework necessitates several key elements to effectively govern high-risk AI systems.
Mandatory risk assessment procedures must be established to identify potential harms, evaluate their likelihood, and propose mitigation strategies. Transparency and explainability are crucial, requiring developers to provide clear explanations of how their systems function, enabling users to understand and challenge AI-driven decisions.
Strict data governance policies are essential to ensure AI systems are trained on high-quality, representative, and unbiased data through clear guidelines for data collection, storage, and use.
Human oversight must be maintained to prevent errors and ensure ethical considerations, potentially requiring human review of AI decisions in specific contexts. Accountability and redress mechanisms are necessary to establish responsibility for harm caused by AI systems, providing remedies for victims and holding developers and deployers accountable.
Finally, investment in training and education programs is vital to build expertise in AI ethics, risk management, and regulation, ensuring Zimbabwe possesses the skills and knowledge to effectively govern AI technologies.
Learning from others
While we can learn from international examples like the EU AI Act, Zimbabwe must develop its own AI policy framework that reflects its unique values, priorities, and challenges. We must avoid the temptation to simply copy and paste existing regulations without critically examining their suitability for our local context.
Instead, we should draw on our own rich history of community engagement, participatory decision-making, and respect for human rights to create an AI policy that is both innovative and equitable.
The task ahead is not easy. Navigating the complex landscape of AI risk requires careful consideration, open dialogue, and a commitment to putting the interests of all Zimbabweans first.
But by embracing a nuanced understanding of high-risk AI as the cornerstone of our National AI Policy, we can pave the way for a future where AI serves as a powerful engine for inclusive and sustainable development. To achieve all this, we must put understanding high-risk AI at the centre of the new National AI Policy.
About the Author: Dr Evans Sagomba is a Doctor of Philosophy and Chartered Marketer (CMktr, FCIM) with an MPhil and PhD in Philosophy. He specialises in AI, Ethics, and Policy Research, and is an AI Governance and Policy Consultant. His expertise extends to the Ethics of War and Peace, Philosophy of Development, and Political Philosophy. [email protected]. ORCID: 0009-0007-0681-0329. Social media handles; LinkedIn; @ Dr. Evans Sagomba (MSc Marketing)(FCIM )(MPhil) (PhD) X: @esagomba
Online Reporter