What makes good AI governance?
In the third edition of the DSA AI Webinar series, we discussed this question in depth and realised that governance is more than rules and policies.
In edition three, our guest was Dr Tabani Moyo, who has a wealth of experience in artificial intelligence, freedom of expression and advocacy.
The third edition’s theme was Smart and Safeguarded: AI Ethics and Governance for African CSOs. The third edition consolidated the takeaways from the first two sessions on gender bias and privacy.
As AI acceptance sweeps across the continent, it is paramount that we focus not only on how to use it effectively but also on how to govern its use. Across the globe, and more specifically in Africa, conversations about governance at the state level have continued, with many African countries adopting AI strategies. At the local level, many organisations, newsrooms and civil society organisations are beginning to consider adopting their own strategies and governance frameworks.
As we discussed AI governance, Dr Tabani emphasised that governance has shifted from merely cyber security to more comprehensive approaches. This means that governance no longer merely looks at potential harms but also fully integrates with law, thereby making AI governance strategies more holistic.
As AI rapidly proliferates in our everyday lives, it can be hard to work out what to prioritise when it comes to governance. While some people may prefer strategy over law, it is important that both are prioritised. However, strategy should come before implementation, because a clear road map is needed to know where you intend to go.
AI is a global innovation, and as governance and strategy are developed to match rising
global demands, Dr Tabani noted that innovations need to focus on the realities of the
people drafting the strategy. For example, in the West, AI strategies focus heavily on health care or judicial services. However, Africa largely supports agro-based economies, and innovations should reflect this. To this end, we need to continue investing in indigenous knowledge systems and use them as catalysts for growth and digital innovation across the continent.
We therefore need to ensure that the anchor sectors of the economy are at the forefront of adoption, development and use of AI to stimulate societal transformation and development.
However, one clear issue stood out: many of our knowledge systems and records are still kept in physical form. This ties in with digital colonialism. When Africans are only end users, their knowledge, data and experiences are extracted from them without any control or oversight. Thus, Dr Tabani called for young people to be keen to promote local solutions and actively participate in the local economy. We need to be actively involved in creating and using local solutions.
So, we asked the question – what are the areas CSOs need to be wary of when drafting their AI governance strategies? Dr Tabani wrapped it up nicely:
1. Regulate what you know: Organisations need to fully understand the technology they are working with and whether it truly suits them.
2. Avoid copy-and-paste strategies: Different organisations have different needs. What may have worked for someone else may not suit you. Create a strategy that suits you.
3. AI is not cheap: As organisations develop a policy and framework, ensure that the
framework you have created is one your organisation can afford. Free access to AI
tools leads you to only reaping limited benefits.