The gap between
what African states
declare about AI
and what they can
actually do.
A diagnostic for AI sovereignty and readiness across nineteen countries. Two dimensions, eleven sub-dimensions, forty-four indicators, 836 data points. From the working paper Negotiating Intelligence: African Agency in the Global AI Value Chain.
Score your countryThe global AI economy is shaped by two trends moving in opposite directions. Frontier AI training is concentrating in a small number of US and Chinese firms, with single training runs projected to cost between 10 and 100 billion dollars by 2030. At the same time, the cost of running an existing capability is falling by roughly three to five times each year. The strategic question for African states is not whether they can match frontier compute spending. It is whether they will be ready to govern, procure, and deliver the capability that diffuses from that frontier. The Sovereignty Plane measures that readiness as a substantive matter, not a formal one.
A diagnostic
for the gap.
Most analyses of African AI governance count strategies, laws, and signed declarations. This is a two-dimensional instrument for measuring whether the institutions behind those declarations can actually do anything — applied to nineteen countries (fifteen African states plus four reference anchors) across forty-four indicators.
Four quadrants, one coherence line.
The plane is a diagnostic, not a verdict. What matters is not where a state currently sits but the direction of its movement — toward the diagonal where declaration and capacity coincide, or away from it.
Sovereignty Theatre
Strategies, laws, and institutions accumulate faster than the capacity to enforce, evaluate, or renegotiate. Seven of fifteen African states sit here.
Negotiated Interdependence
Formal architecture matched by operational capacity. The state can enforce against major providers and bargain credibly. Only South Africa among the African sample.
Ad-Hoc Capability
Capacity outruns formal architecture. Ecosystems, private investment, or diaspora networks generate capability without state-led planning. Tunisia is the only African case.
Dependency by Default
Limited formal architecture and limited substantive capacity. The state is a passive recipient of AI systems built, governed, and operated elsewhere.
Writing at the intersection of governance, capacity, and care.
Negotiating Intelligence: African Agency in the Global AI Value Chain
Prepared for submission to Data & Policy (Cambridge University Press) · April 2026Introduces the formal–substantive sovereignty distinction and the Sovereignty Plane. Applied to fifteen African states and four reference countries across forty-four indicators and 836 scored data points (660 across the African sample, 176 across the four reference anchors).
Cost-consequence analysis of a digital health–enabled NCD management intervention in Ghana
Springer Nature · Cost Effectiveness and Resource Allocation · 2026 · open accessRetrospective economic evaluation across sixteen health facilities and 705 participants. The intervention was cost-saving (~$21,000 annual savings) with superior clinical outcomes — a dominant economic profile.
Negotiating Losses in the Age of Digital Transformation: The Case for Augmented Human Intelligence
LinkedIn · 2025Explores what societies quietly surrender in the pursuit of digital progress — from algorithmic curation of memory to the hidden costs of cognitive augmentation — and argues for reclaiming human editorial agency over technology.
A Tale of a Miner's Companion: Like a Canary in a Coal Mine
LinkedIn · 2020Draws a parallel between the canary in a coal mine and AI-enabled diagnostics in resource-limited health systems, drawing on field experience deploying digital X-ray with AI algorithms in a Sierra Leone hospital.
Cost-effectiveness analysis of digital health interventions for NCD management in Ghana
KNUST · Kwame Nkrumah University of Science and Technology · 2024Built cost-effectiveness models using real-world programme data from LMIC settings. Obtained ethical approval, managed multi-facility data collection, and conducted quantitative analysis using standard health economic evaluation frameworks.
Ten years inside African digital transformation.
The Sovereignty Plane is a theoretical instrument. It was written because the practitioner experience below made the gap impossible to unsee.
- 2025 — now
Tony Blair Institute for Global Change
Manager, Technology and Innovation
- Aug — Dec 2025
Sand Technologies
Consultant — National Health Information Hub
- 2023 — 2025
Senior Programmes Lead
- 2021 — 2022
Senior Technical Advisor — Digital Transformation & PPP
- 2016 — 2021
Project Manager — Africa
I'm a digital policy practitioner and health systems researcher based between Freetown and Accra. For the last ten years I have worked inside the machinery of African digital transformation—running multi-country digital health programmes, managing innovation funds, drafting national strategies, advising ministries. I've seen what it looks like when a state signs an AI strategy and what it looks like when it actually has to enforce one. Those are not the same thing.
The Sovereignty Plane is a theoretical framework I wrote because the practitioner experience made the gap impossible to unsee. The Global AI Majority research community has converged on what needs to be built—sovereign AI safety capacity, inclusive participation, Africa-centric evaluations. It has not yet agreed on how to measure whether it is being built. The Plane is my contribution to that measurement problem.
I hold an MPhil in Health Systems Research and Management from KNUST, am PMP-certified and HL7 FHIR-certified, and serve on the steering committee of Health Technology Assessment International’s Developing Countries Interest Group. I’m an Acumen Academy Fellow.
If you’re working on Global AI Majority participation, Africa-centric safety evaluation, or the measurement of sovereign capacity in frontier AI governance—whether from a research, policy, or funder perspective—I’d welcome a conversation. I’m particularly interested in collaborations that test the Plane against new data or extend it to other regions of the Global AI Majority.
The best way to reach me is directly. appiahk4@gmail.com
Also reachable on +232 80 266 124