What evidence will help our case?

AI Governance for Africa > Part 3 > Section 4

Evidence is crucial in any advocacy strategy. In building out your advocacy strategy, you need to take some time to think about who you need to persuade, what evidence you need to persuade them, and what you need to do to get it.

---

Evidence is crucial in any advocacy strategy.^[1] Any effort to shape AI regulation based on principles of transparency, accountability, inclusion, and minimising harms to human rights, needs to include compelling evidence for its case.

The need is especially clear given the tendency for African governments’ rhetoric on AI policy to be bullish on the opportunities for economic growth and ‘magic fix’ technological solutions, often glossing over any of the attendant risks to human rights and social inclusion.

On the risk side, there is a growing body of evidence globally of the potential harms of unregulated AI. But as with many aspects of AI knowledge, the evidence of AI harms is often skewed to the Global North with many of the best documented examples and systematic studies occurring in North America and Europe. While evidence from other countries is an important part of making the argument, this suggests the need to build better bodies of knowledge of how AI is being deployed in African contexts.

On the implementation side, there are several new and emerging examples of AI regulation and governance frameworks across the world. This kind of evidence can help local policymakers too, and can be used by advocates to create a benchmark for best practice.

Data mapping exercise

Advocates for AI regulation can map out their existing evidence base, and their existing data needs, using a simple data mapping exercise. For example:

Key advocacy message	Key actors to persuade	Data needed to persuade key actors	What data do you already have?	What data do you still need?
“AI technologies need built-in transparency tools so users can understand what decisions are being made.”	Policymakers, technologists	Case studies of how lack of transparency leads to unfair outcomes	Research publication on AI bias in the United States	Local case studies of AI bias in my country
“AI governance is achievable – we can build on global best practice, and become a global leader on this!”	Policymakers, political leaders	Comparative policy	UNESCO Recommendations on Ethics in AI	Comparative review of AI governance instruments

Key advocacy message	Key actors to persuade	Data needed to persuade key actors	What data do you already have?	What data do you still need?
“AI technologies need built-in transparency tools so users can understand what decisions are being made.”	Policymakers, technologists	Case studies of how lack of transparency leads to unfair outcomes	Research publication on AI bias in the United States	Local case studies of AI bias in my country
“AI governance is achievable – we can build on global best practice, and become a global leader on this!”	Policymakers, political leaders	Comparative policy	UNESCO Recommendations on Ethics in AI	Comparative review of AI governance instruments

---

Navigate sections

Explore the rest of the toolkit