IGF 2023 WS #394 How Governments can harness data to develop AI systems

Organizer 1: Jackline Akello, 🔒
Organizer 2: Samantha Msipa, Research ICT Africa
Organizer 3: Sandra Makumbirofa, Research ICT Africa

Speaker 1: Jackline Akello, Civil Society, African Group
Speaker 2: Samantha Msipa, Civil Society, African Group
Speaker 3: Sandra Makumbirofa, Private Sector, African Group


Sandra Makumbirofa, Private Sector, African Group

Online Moderator

Sandra Makumbirofa, Private Sector, African Group


Samantha Msipa, Civil Society, African Group


Panel - 60 Min

Policy Question(s)

How are national and international laws applied in cyberspace in the context of digital sovereignty?
How do current cross-border data flow frameworks address legitimate developmental needs in relation to AI?
What role can data and AI play in tackling sustainability issues such as climate change, biodiversity, conservation, and water scarcity?

What will participants gain from attending this session? Participants will:
Get to understand the critical role that governments can play as providers of data for AI/ Different ways governments can use the data they steward to support the development of AI.
They will get to understand the mechanisms that can be put in place to enable data sharing and cross-border data flows (and also data stewardship). This knowledge will help them navigate the data ecosystem and make informed decisions regarding AI development.
Exposure to/ understanding of the type of data held by governments i.e., administrative data, statutory registers, or routine monitoring data - and their potential to support the development of AI systems.
It will enable them to gain insight into the regulatory environment (policies, initiatives, and frameworks) for governing AI-relevant data in the hands of the government. This provides regulatory tools which participants looking to collaborate or work within the framework set by the governments can use.


This session will discuss the data ecosystem in Africa within the context of developing AI systems. The focus will be on analyzing the continent's data processing activities to demonstrate how a strong data ecosystem can harness the potential benefits of AI. Additionally, the session will explore how a robust data-sharing environment and framework can promote AI-driven economic growth in Africa, emphasizing the regulatory measures needed to facilitate data-sharing.

The discussion is prompted by the limited availability of data and the lack of quality and timeliness of data in Africa for AI system development. The continent faces a "data desert" where there is insufficient data to train AI algorithms. Contrary to the perception of a data desert, Africa possesses a diverse data ecosystem primarily data generated by government institutions. While governments generate substantial data that could be used for AI system development, there are issues with data quality and accuracy. Furthermore, government-backed open data initiatives often struggle to progress beyond their launch due to inadequate updates and resource allocation.

Leveraging the available data resources can empower the creation of impactful AI systems that address Africa's most pressing challenges without exacerbating societal and cultural norms that marginalize vulnerable populations, including women. By establishing a robust data ecosystem that facilitates data sharing and cross-border flows, AI algorithms can be trained on data that specifically tackle Africa's unique issues and encompasses disadvantaged groups.

The discussion will primarily focus on examining the African data ecosystem to devise legal, infrastructural, and structural frameworks necessary to establish a strong data ecosystem in Africa. This ecosystem should enable AI to mitigate harm to marginalized groups and foster economic growth across African countries.

Expected Outcomes

We shall start by publishing a blog post on the issue. Use the blog to invite those interested in the issues for a panel discussion. This will help us identy gaps in the policy space from a continental level.

Hybrid Format: We shall use a Virtual Event Platform such as Zoom and provide interactive features such as a Mentimeter, and live Q&A sessions that can enable us to actively engage with onsite and online attendees. While one person speaks, the other will have a follow-up pop-up quiz to gauge whether the audience is following the discussion clearly. The moderator will ensure that speakers are adhering to the timeframe and there is ample time for engaging with the audience.