Mauritius’ new AI policy makes ethics mandatory, not optional

While many African countries race to deploy artificial intelligence, Mauritius has made governance and ethics the starting point of its AI strategy, rather than a problem to solve after the technology is in use.

Central to the strategy is the FAIR framework, a set of guidelines that governs how AI systems are designed, deployed, and managed. It sets clear expectations across sectors and applies to the entire AI lifecycle, from design and development to deployment, monitoring, and eventual decommissioning.

Mauritius’s approach reflects a broader shift in how African countries may position themselves in the AI landscape. While larger markets such as Nigeria and Kenya emphasise scale and ecosystem growth, and South Africa focuses on institutional regulation, Mauritius is advancing a governance-led model centred on enforceable standards.

The Mauritius National AI Strategy 2025–2029, alongside the FAIR Guidelines introduced in April 2026, is designed to be vendor-neutral and border-agnostic. Any AI system operating within the country, regardless of origin, must comply with a unified set of ethical and operational standards.

Imported AI tools are subject to the same level of scrutiny as domestic systems. The framework requires compliance with principles of fairness, accountability, inclusiveness, integrity, and responsibility. In high-risk sectors such as fintech and gaming, systems must undergo bias audits to mitigate discriminatory outcomes. Accountability provisions also require foreign providers to designate locally based representatives who can be held responsible for system outcomes.

Any AI system that affects individuals, organisations, or public interests in Mauritius falls within the framework’s scope, reflecting a recognition that AI risks are not bound by geography and that governance should be determined by impact rather than origin.

Although the FAIR Guidelines are currently non-binding, there are no immediate legal penalties or fines for non-compliance—at least not yet; they are designed with a clear legal and policy trajectory. They are expected to shape government policy, inform sector-specific regulations, influence procurement standards, and eventually underpin future legislation.

In effect, Mauritius is building a regulatory framework that can evolve alongside the technology, rather than locking in rigid rules too early. This contrasts with approaches like South Africa’s Draft National AI Policy, which proposes steep penalties—including fines of about $530,000 or up to 10 years in prison—for serious ethical breaches.

The Mauritius approach allows the country to remain flexible while still establishing a stable reference point for accountability. Policymakers, regulators, businesses, and even courts can rely on these principles as AI adoption expands.

The framework has four pillars: fairness, accountability, inclusiveness, and integrity. Each addresses a specific risk that has emerged in global AI deployment and is tied to concrete expectations.

Fairness focuses on preventing bias. AI systems must not discriminate based on income, gender, ethnicity, or geography, the policy stated. This is particularly important in a small and diverse society, where flawed systems could quickly exclude entire groups from access to services or opportunities. To address this, the guidelines emphasise the use of representative local datasets and require bias testing, especially in high-impact sectors such as finance and public services.

Accountability tackles one of AI’s most persistent challenges: the “black box” problem. Under the FAIR framework, there must always be a clearly identifiable party responsible for an AI system’s decisions. This includes defining liability, maintaining audit trails, and establishing mechanisms for redress when harm occurs. AI decisions are not meant to be opaque or unchallengeable.

Inclusiveness ensures that the benefits of AI are widely distributed. Rather than concentrating advantages among large firms or urban populations, the strategy promotes AI literacy through initiatives like “AI for All,” supports small and medium-sized enterprises, and expands access to digital infrastructure. The goal is to prevent a new form of inequality—what the policy’s authors describe as a potential “digital divide 2.0.”

The final pillar, integrity and responsibility, addresses the technical and ethical robustness of AI systems. It covers data governance, privacy, cybersecurity, and safeguards against misuse, including fraud and manipulation. For a government that plans to integrate AI into public service delivery, trust in system reliability is essential.

What sets Mauritius apart is not just the inclusion of these principles, but how they are embedded into the broader economic strategy. The FAIR framework is tied directly to procurement decisions, system design, and policy development. It is positioned as a baseline requirement, not optional guidance.

This reflects a broader strategic choice: as a small, open economy of just 1.26 million people and a roughly $15 billion GDP, Mauritius cannot compete on scale with larger economies like South Africa, with an over $400 billion GDP.

It is not that South Africa and Nigeria are ignoring trust. The difference lies in priorities and timing. Mauritius is using its smaller size to position itself as a focused, “boutique” AI regulator, while South Africa and Nigeria must balance building trust with driving the scale of growth their larger economies demand.

In doing so, it hopes to attract investment, build partnerships, and integrate into global AI value chains.

The country’s economic ambitions reinforce this direction. AI is seen as a new growth pillar, alongside traditional sectors like manufacturing, whose contribution to GDP has steadily declined—from over 20% in the late 1990s to about 10.7% in 2020, and only a modest recovery to roughly 12.8% in 2024.

According to the policy, the country now sees AI as a way to revitalise these sectors, improve efficiency, and create new opportunities in areas such as fintech, logistics, and the ocean economy.

To drive this transformation, Mauritius is building institutional capacity in the form of an AI Council. The council would be supported by public and private sector stakeholders, and international experts, who will oversee implementation, coordinate projects, and measure socio-economic impact. Incentives such as tax credits, grants, and regulatory support are also being deployed to encourage adoption.

This governance-led approach stands in contrast to other African AI strategies. Nigeria, for instance, is prioritising large-scale deployment and talent development, with governance structures still evolving. Kenya is focused on building a regional innovation hub and a powerful AI sheriff, while South Africa is leaning toward a more regulation-heavy model with multiple oversight bodies.