Portfolio 03
AI Governance & Accountability.
Documenting AI harms in Africa, and helping civil society respond to them.
The gap
Who responds when an AI system causes harm?
AI is entering African public services, security systems and financial institutions faster than the safeguards around many of those deployments are developing.
Much of the current conversation focuses on frameworks, principles and regulation. Far less attention is given to the operational question: what happens when an AI system causes harm, and who has the capacity to document, investigate and respond?
A security perspective
We are building that capability.
01
Detect
02
Document
03
Investigate
04
Respond
Our thinking
Three positions we hold.
01
Governance needs evidence
Policies developed without a meaningful record of real-world incidents in African contexts risk becoming borrowed frameworks rather than responses to local experience. The evidence base is still thin.
02
AI is becoming part of the threat landscape
Civil society organisations are already confronting AI-enabled surveillance, impersonation, disinformation and targeted attacks. These are not only governance questions; they are security problems.
03
Security institutions require scrutiny too
The introduction of AI into policing, intelligence, border management and other security systems creates a significant but comparatively under-examined risk surface. These deployments deserve the same evidentiary scrutiny applied to other powerful technologies.
Our approach
Four lines of work.
Each one extends a capability we already run for civil society into AI-enabled harm.
Step 01
Build an Africa AI Incident Observatory
We are developing an observatory to document and analyse AI-related harms affecting African communities and civil society, creating an evidence base that can inform researchers, advocates and policymakers.
Step 02
Extend security operations to AI-enabled threats
Our existing incident-response and security operations capabilities provide a foundation for helping civil society understand and respond to emerging AI-enabled attacks.
Step 03
Investigate high-impact deployments
Where AI systems create significant consequences for rights, safety or civic participation, we intend to apply the same investigative discipline used in our spyware work: establish what happened, preserve evidence, identify affected parties and document the chain of responsibility.
Step 04
Bring security into AI governance
We are contributing a civil society and security perspective to emerging African AI governance processes, with particular attention to accountability, institutional safeguards and high-risk government deployments.
Where we are
2026
Foundation year
2026 is the foundation year for this portfolio. We are building internal expertise, completing advanced AI security training, studying existing governance frameworks and designing the Observatory before expanding its public outputs.
Contribution
Nigeria’s proposed National AI Trust
We contributed to the co-creation of Nigeria’s proposed National AI Trust, including work on its legal vehicle, civil society representation and the treatment of the security sector as a distinct AI risk area.
The objective is to build the evidence, investigative capability and civil society infrastructure needed to make AI governance accountable in practice.
Objective
Partnered with
- West Africa Civil Society Institute (WACSI)
- The Hague Institute for Innovation of Law (HiiL)
- Nigeria Media Innovation Program (NAMIP) Media Development Investment Fund
- DefendDefenders
- Digital Society of Africa (DSA)
- Open Technology Fund Funder
- TechSoup