Nigerian ministries, departments, and agencies are buying AI tools with no rulebook, no shared lessons, and no paper trail when things go wrong.
A new U.S. Government Accountability Office report, released April 13, 2026, exposes the same structural failure in Washington. U.S. federal agencies roughly doubled their AI footprint between 2023 and 2024. Not one of the four agencies the GAO reviewed had internal rules requiring staff to record what went wrong in those procurements. The knowledge is closed with each contract.
The Federal Capital Territory has the same problem and fewer resources to absorb it.
Nigeria's AI Procurement Has No Rulebook Yet
Less than 10% of Nigerian government operations have been digitized, according to NITDA Director-General Kashifu Inuwa. The MDAs buying AI tools now have no procurement infrastructure under them. No standard contract terms. No data-rights clauses. No testing benchmarks.
The GAO flagged the same gaps in the U.S. context. Agencies couldn't find specialists to interrogate vendor proposals. They couldn't account for what AI was actually costing them. In Nigeria, both problems are more acute.
There is no centralized body collecting and redistributing procurement lessons across MDAs. Every agency that buys an AI product starts from scratch. [Internal link:
Anambra State's ICT agency built an AI model that flagged a ₦2.1 billion unspent education allocation and caught a vendor awarded five contracts in three weeks. That is a concrete win. But the architecture for sharing it so Kogi, Bauchi, or Imo can replicate the result without repeating Anambra's early mistakes does not exist at the federal level.
What the U.S. Got Wrong And What Nigeria Must Avoid
The GAO reviewed AI acquisitions across four agencies: Defense, Homeland Security, GSA, and Veterans Affairs. None required staff to document procurement lessons. None shared knowledge with other agencies.
GAO issued four recommendations, one per agency, requiring each to build a formal lessons-capture process and submit findings to a GSA-managed repository.
Nigeria has published its National AI Strategy and is advancing the National Digital Economy and E-Governance Bill. That legislation needs a procurement knowledge-sharing mandate built into it, not bolted on later.
The GAO mapped three fault lines in AI buying: who drives the decision, the agency or the vendor; what legal instrument governs the deal; and whether the agency owns the AI or rents its outputs indefinitely. Nigerian MDAs face all three. They have fewer technical staff to navigate them and fewer legal guardrails protecting public funds.
The Fix Nigeria Needs Right Now
The National Digital Economy and E-Governance Bill must require every MDA that procures an AI system to document the process, vendor selection criteria, contract terms, data-rights agreements, what worked, and what failed, and submit it to a NITDA-managed repository. That clause is not optional. Without it, the bill is infrastructure with no plumbing.
Nigeria's Digital Sovereignty and Fair Data Compensation Bill, 2025, would require foreign companies using Nigerian data to train AI models in Nigeria and contribute 2% of their annual revenue in Nigeria to a Nigerian AI development fund.
70% of Nigeria's online population uses generative AI tools, according to NITDA Director-General Kashifu Inuwa, well above the global average of 48%. Nigeria's government is still working out how to buy it responsibly.
Nigeria cannot afford another budget cycle of AI procurement without a paper trail. The GAO's warning costs U.S. agencies nothing to ignore. Nigeria does not have that luxury.
Stay ahead of every tech story shaping Nigeria and the world. Sign up for the GizPulse newsletter sharp, fast, and built for people who need to know first.
