Africa Innovation Portfolio

At the end of the dialogue - there must be consensus on how all the countries will make sure that the most vulnerable will not be left behind as AI advances continue to occur (some without their awareness) The must be clear outlines of how AI sovereignty will be gorvened. The risks (especially) unknow risk of agentic AI can be exposed in order to ensure security of states are not compromised.

• Protection and promotion of human rights
• Open-source software, open data and open AI models
• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI

AI has a potential to widen the digital divide as well as a potential to impact the most vulnerable in terms of economic participation, therefore, as AI advances, considerations on how livelihoods can be maintained in the advent of loss of eligibility to participate in the economy. Safety and Security is at risks across all spheres of society as long as we cannot have full transparency of the logic of the AI systems that are implemented by large corporates to make high impact decisions such as hiring, financial access, etc. Accessibility to open source systems will enable capacity building, and reduce dependency on ready made solutions. Protection of human rights remains a critical element with the ability of AI to generate fake content and material which can have detrimental impact on peoples well being.

With the emergence of the recent wars, we need to highlight the role of AI in peddling misinformation and propaganda. Secondly, I want to highlight that we use lessons from the implementation of the SDG2030 to ensure that we all guard the speed at which we drive AI dialogue and any actions out of it. So that we don't come to the same pitfalls encountered with driving SDGs.

Social, economic, ethical, cultural, linguistic and technical implications of AI: Challenges: There is a gap for exploitation as people do not fully understand, what it means to train models, and what the acquisition and ownership of indigenous knowledge by foreign systems that can potentially gain from the use of the information. In my country, we currently experience a high unemployment rate, and the adoption of AI does impose a threat to the marginalised, and most vulnerable participants in the economy. Opportunity: There is opportunity to upskill the workforce and the unemployed to be economically prepared for AI impacts on their livelihoods. Another opportunity exists in empowering the small enterprises to be enablers of AI awareness especially in the most underserved communities.

The AI dialogue can help bring various communities to engage, promote participation towards clarifying the AI blind-spots, it can also help facilitate cross-skilling, and learning from other's mistakes/lessons learnt in the process of defining policies, adoption, implementation, and governance.

I think the AI Dialogue must feature in initiatives around sustainability and world peace and security, as the AI impacts touch on these area quite significantly, and must not be an after thought. I think in the G20, G7, WEF and in the UN Security Council there must be some connections built. This will ensure that world leaders engage with AI governance concerns at a leadership level.

Open engagement sessions, such as conference/summit - with groups sessions Open call for written submissions to particular question(s) Insights and show and tell sessions Polls/Surveys

Youth, women, the blind, deaf, and people in rural communities who may not necessarily speak English (as most AI info is in English)

Apply the Nominal Group Technique to gather important actions and to prioritise them.

An approach that could promote effective AI governance is implementing ethical usage standards or guidelines e.g. in sectors such as education where plagiarism is not acceptable, in workplaces where authenticity is required and there is interest in assessing the impacts of AI in productivity; transparency of methodology or algorithm used in sectors where evaluations affecting personal outcomes such as recruitment, finance applications. Where AI is being used, those who use it must know the extent of acceptable usage, and where people may not be aware it is used, there must be disclaimers in understandable language of what the usage means and how it impacts them. Thank you.