OpenUp believes that technology should strengthen democratic participation, not replace it. This is why we have previously discussed AI's role in augmenting participation. Bases on our continuing work on AI and technology in democratic space, we are sharing today our latest policy brief: Responsible Use of AI in the South African Public Sector. This knowledge has been generated from a real-world project that tested how artificial intelligence can meaningfully support citizen engagement.
From Technical Experiment to Policy Insight
In 2025, we collaborated with the Policy Innovation Lab at Stellenbosch University and the GIZ "Data2Policy" project to design and review (technically and ethically) a potential Citizen Generated Data Tool that incorporated generative AI. In undertaking that work, this policy brief emerged as a way of documenting our broader findings to feed into the responsible technology community for the public sector (and help inform the emerging development conversations on Digital Public Infrastructure). The policy brief seeks to grapple with the particular challenges of deploying AI within South Africa's political context. We examined case studies from municipalities using algorithmic decision-making, analysed capacity constraints across government departments, and mapped the governance frameworks that shape responsible AI adoption.
What became clear is that technical relevance alone isn't enough. AI systems serving the public sector need to be designed with transparency, explainability, and data portability as core features, not afterthoughts. They need to avoid vendor lock-in. They need to complement - not replace - the in-person engagement that builds political trust. And they need clear accountability mechanisms for when things go wrong.
Five Critical Dimensions
Our analysis identified five dimensions that should guide any AI deployment in the public service: the capacity to design and adopt these systems sustainably; the realities of long-term maintenance; the importance of maintaining personable participation; ensuring accountability and responsibility; and above all, the pursuit of genuine public purpose.
That last point deserves emphasis. We believe AI should advance human flourishing and strengthen democratic processes - not simply pursue efficiency, or even just technology, for its own sake. Every technical decision should be measured against whether it genuinely serves citizens and enhances public participation.
Practical Recommendations for All Stakeholders
The brief concludes with concrete recommendations for policymakers, developers and vendors, and government departments. The recommendations are grounded in our lived experience building civic technology in the South African context, where digital inequalities, limited technical capacity, and vendor dependencies create real constraints.
We invite you to read the full policy brief. As South Africa's public sector stands at this critical juncture in AI adoption, we believe the technical community has a responsibility to build thoughtfully, transparently, and always in service of democratic participation.
Read the full policy brief below .
This project was implemented by Open Up, on behalf of the GIZ project “Data2Policy” commissioned by the Federal Ministry for Economic Cooperation and Development (BMZ).
