Nigeria's AI Governance Score Rose From 7 to 46 in Two Years. The Real Story Is What Still Doesn't Count.

Nigeria jumped 42 places to become Africa's top-ranked country in the 2026 Global Index on Responsible AI, but the index's own authors warn that scoring a policy is not the same as protecting a citizen from it.

Forty-five point nine three. That is Nigeria's score on the second edition of the Global Index on Responsible AI (GIRAI), published this month by the Cape Town-based Global Center on AI Governance. Two years ago, on the same 100-point scale, Nigeria scored 7.21. The country has climbed from 80th to 38th globally, overtaking Egypt and Kenya to become the highest-ranked nation on the continent -- and the index's authors have flagged it as one of a handful of global \"Bright Spots.\"\n\nFor a government trying to convince investors that Nigeria is a serious place to build AI, that is a genuinely good number to be able to cite. Bosun Tijani, the minister of communications, innovation and digital economy, wasted no time doing so. \"This recognition is a testament to Nigeria's deliberate efforts to build an AI ecosystem that is inclusive, responsible, and aligned with our development priorities,\" he said in a statement issued through his office on July 9, framing the ranking as support for President Bola Ahmed Tinubu's ambition of a $1 trillion economy.\n\nBut read past the press release, into the index itself, and a more complicated picture emerges -- one that Himilo Post's African tech readers, many of them building the products this framework is meant to govern, should sit with before celebrating.\n\nWhat actually moved\n\nGIRAI scores 135 countries across five dimensions: inclusion and diversity, ethics and sustainability, labour and skills, trust and safety, and AI use in public services. Sixty percent of each dimension's score comes from documented government policy and initiatives; 30% from structural \"enabling conditions\" like infrastructure and skills base; and 10% from civil-society engagement. A separate penalty -- the Unacceptable Risk AI Systems deduction -- subtracts points for documented state misuse, from biometric surveillance to AI-enabled disinformation campaigns.\n\nNigeria's rise traces to specific, checkable government output: the National Artificial Intelligence Strategy, which mandates AI-literacy programmes and teacher training; the 3 Million Technical Talent (3MTT) programme, which has run structured AI and machine-learning cohorts nationwide since 2023; and the Nigeria Data Protection Act together with last year's General Application and Implementation Directive, which bars decisions about children being made by automated systems alone and requires parental consent for handling their data. Each of those is a real, auditable policy instrument, not a slogan -- which is precisely why GIRAI credited Nigeria's climb to them rather than to rhetoric.\n\nNigeria also climbed 31 places (103rd to 72nd) in the separate Oxford Insights Government AI Readiness Index in January, a second, independently constructed measure pointing the same direction. Two different indices, built by two different institutions on two different methodologies, telling a consistent story is a stronger signal than either alone.\n\nWhy the global average matters more than the ranking\n\nHere is the number that should worry anyone reading Nigeria's 38th place as \"mission accomplished\": the average GIRAI score across all 135 countries is roughly 35 out of 100. Nigeria, at 45.93, is now comfortably above that average -- but the average itself is grim. Global North countries average 55; the Global South averages 27. A high-45s score puts Nigeria ahead of most peers in a low bar-clearing exercise, not at any meaningful standard of AI safety.\n\nRobert Opp, UNDP's chief digital officer, put the sharper point on it in the report's foreword: where responsible-AI frameworks exist, only 55% of cases show evidence they are actually being implemented -- and that figure drops to 45% in the Global South, Nigeria's own peer group. Fewer than one in five of the 135 countries assessed require their own governments to disclose when they are using an algorithmic system on citizens. Only 28 have an independent oversight body watching how AI is deployed in the public sector. \"Implementation gaps, in many contexts, are development gaps,\" Opp wrote -- a warning aimed squarely at governments tempted to treat a published strategy document as the finish line rather than the starting gun.\n\nThe framework-versus-enforcement gap, applied to Nigeria\n\nNigeria's own gains sit inside that caveat. The Data Protection Act and its 2025 directive are real legal instruments with real restrictions written into them -- but a law's existence and a law's enforcement are different facts, and GIRAI's broader finding is that only 22% of Global South framework provisions are binding, against 58% in the Global North. The report does not allege Nigeria's specific frameworks are toothless; it does note, at the systemic level, that most Global South responsible-AI provisions are not.\n\nThat distinction matters for the founders, engineers and operators who make up Himilo Post's core audience, because a country's AI-governance score increasingly functions as a signal to capital -- the same way a sovereign credit rating does. Investors underwriting an AI deal in Lagos will read \"Africa's top-ranked country in responsible AI\" as a genuine positive: it lowers the perceived regulatory-surprise risk of building there relative to markets without any comparable framework at all. But the index's own authors are telling those same investors, in the same document, not to mistake the existence of a framework for evidence that it is enforced, funded, or staffed. A minister citing a global ranking and a regulator actually auditing an automated-decision system against the GAID directive are two different stages of the same process, and only Nigeria's own follow-through will show which stage the country is actually in a year from now.\n\nWhat to watch\n\nThe honest read of this ranking is neither dismissal nor celebration. Nigeria did something real: it built named institutions, passed enforceable-on-paper protections for children's data, and ran a talent programme at national scale, and an independent, multi-country index credited all three specifically. The open question the index itself raises is whether the 3MTT graduates, the NAIS mandates and the GAID's automated-decision restrictions will show up, a year from now, in the kind of implementation evidence that moved Nigeria's enabling-conditions score -- or whether the 45.93 becomes a number the government cites at conferences while the oversight bodies, disclosure requirements and enforcement staff that would make it real remain unbuilt. GIRAI's next edition, due in roughly two years on the same methodology, is the mechanism that will answer that.