The 2026 National E-Governance Bill: What Nigeria Is and Isn't Getting Right

nigeria's public sector is quietly undergoing an algorithmic transformation. from identity matching at the national identity management commission (NIMC) to automated fraud detection in public procurement, ai systems are increasingly embedded in the machinery of state. yet, until now, these deployments have occurred in a legislative vacuum. the proposed 2026 national e-governance bill attempts to fill this void.

the bill is ambitious. it seeks to create a unified framework for digital public infrastructure, data interoperability, and the procurement and deployment of artificial intelligence by government agencies. but a close reading of the draft reveals significant structural friction between what it aims to do and the actual mechanics of ai governance.

accountability in government ai

the strongest aspect of the draft bill is its stance on public sector transparency. it proposes a mandatory public registry of ai systems deployed by federal agencies—a direct response to civil society concerns about black-box decision-making. if an agency uses an automated system to determine eligibility for a federal loan program or to flag tax anomalies, that system must be logged.

furthermore, the bill requires human-in-the-loop oversight for algorithmic decisions that materially affect a citizen's legal rights. it explicitly outlaws fully autonomous systems for the revocation of licenses or the denial of essential public services. this is a progressive step, drawing obvious inspiration from the eu ai act's "high-risk" classification.

the missing private sector

however, the bill's fatal flaw lies in its jurisdiction. it is strictly an e-governance bill; it applies only to public institutions and government contractors. the massive private sector ai ecosystem—fintechs using proprietary credit-scoring algorithms, private security firms deploying unregulated facial recognition, e-commerce giants using predictive profiling—remains entirely untouched by this legislation.

nigeria is regulating the buyer while ignoring the builder. by contrast, the eu ai act regulates systems based on their risk category, regardless of whether they are deployed by the state or a startup. by restricting its scope to government use, the nigerian bill leaves the majority of algorithmic harms unaddressed.

the question of liability

the draft is noticeably silent on the issue of liability for ai errors. if an automated system wrongfully denies a citizen's passport application or falsely flags a business for tax evasion, who is legally responsible? the procurement officer? the private contractor who built the system? the algorithm itself?

without clear statutory guidance on liability, aggrieved citizens will be forced to rely on traditional administrative law and negligence torts. these legal avenues are ill-equipped to handle the complexities of machine learning models where causality is notoriously difficult to prove. civil society groups have repeatedly flagged this gap, noting that an unreviewable automated decision is functionally indistinguishable from arbitrary state power.

the role of the NDPC

the bill sensibly leans on the existing nigeria data protection commission (NDPC) rather than creating a new ai-specific regulator. it mandates that any government ai procurement must undergo a data protection impact assessment (DPIA) filed with the NDPC. this leverages an existing enforcement mechanism and avoids regulatory overlap.

however, it is uncertain if the NDPC currently has the technical capacity to audit complex machine learning models across dozens of federal agencies. expanding their mandate to include algorithmic auditing requires a corresponding expansion in unbudgeted technical resources.

the bottom line

the 2026 national e-governance bill is a necessary first step, but it is incomplete. it places reasonable guardrails on government ai use but abdicates responsibility for the private sector. organizations building ai tools in nigeria must understand that this bill is merely the opening round of regulation. the rules of the game are being written right now.