The UN Secretary-General’s Strategy on New Technologies calls for the responsible adoption of artificial intelligence and other frontier technologies — like the Internet of Things, cybersecurity, blockchain, robotics, autonomous vehicles, and nanotechnology — to address the Sustainable Development Goals and the 2030 Agenda. Although AI holds great promise to help us advance global development, it also carries risks of intended and unintended harms. While many of the discussions to date have centered around biases, transparency and accountability, the call for a safe and trustworthy AI as well as responsible adoption demands the inclusion of a human rights-based approach that incorporates safeguards around data privacy and protection.
Countries worldwide are at different stages of designing and implementing AI strategies and policies to seize the opportunities of this technology. In Africa, Kenya, Tunisia, South Africa, Ghana, Uganda are already working to develop data protection and ethics strategies. The critical question now is : Which ethical approaches are relevant in the context of the African continent?
To answer this question, UN Global Pulse, the Ministry of Communications for Ghana and the Data Protection Commission for Ghana, with support from Germany’s GIZ, hosted a session and a subsequent workshop on developing an ethical AI framework in African economies during the 1st African Region Data Protection and Privacy International Conference.
The session addressed the overall ethical issues of AI in Africa, while the workshop focused on the development of an ethical AI framework for Ghana. Both events built on the outcomes of the discussions of a previous event co-hosted by Global Pulse and the UN High Commissioner for Human Rights (OHCHR) during the RightsCon conference in Tunis. Experts on AI ethics and human rights from across the globe — representing the UN and other international organizations, civil society, academia and private companies — shared their diverse perspectives first during the session, and then during the workshop.
The general consensus was that there is a need to develop and implement ethical and regulatory frameworks along with sustainable mechanisms to unlock the availability and value of data to maximize the use of AI while limiting possible harms.
“[We] need more than laws – [we] also need policy and technical and legal experts all working together.”Professor Joe Cannataci, UN Special Rapporteur on the Right to Privacy
Participants noted the urgency to invest in education and the need to reimagine traditional school programmes to create tech savvy lawyers and policy makers, as well as to promote gender equality in education.
“We are happy to have launched our ICT for girls programme!”. “These girls will train other girls in their villages and areas.”Magdalene Apenteng, Ministry of Communications for Ghana
Another point of discussion revolved around finding the right balance between the right to privacy, technological innovation, and governance at country level but also in the context of cross border data flows. The latter comments came out of general concern about foreign economies extracting and benefitting from personal data from countries in Africa, using the lack of, or gaps, in data privacy and protection safeguards and regulations.
“An African approach to privacy and protection is not about personal data, but collective rights. [We need to take into account] [p]eople in slums and others who aren’t normally in the data.”Arthur Gwagwa, Research Fellow, Research ICT Africa, University of Cape Town
During the conversation, the failure to utilize new data sources and artificial intelligence to help achieve the SDGs was deemed as unacceptable.
The non-use of AI where it could help is also unethical. However, any use of AI needs to be done responsibly.”Boris Wojtan, Director of Privacy, GSMA
And having access to data but not having the right policies or tools in place precludes data use for social benefit.
The events in Ghana clearly identified AI as a cross-sectoral subject, with legal, technological, and ethics standards complementing each other in an AI framework to be created at the national level. In the case of Ghana, “we definitely need a strategy, and we are creating one,” concluded Apenteng suggesting that Ghana is embarking on the journey to design an ethical AI strategy, and is doing so, in consultation with experts across organizations and industries.
Here are some key recommendations which emerged:
- Participants stressed the importance of having strong data protection regulation, privacy and ethics-by-design embedded in all AI processes.
- Building capacity and expertise within African economies is one of the key components for ensuring that AI is implemented safely and sustainably.
- Participants agreed on the need to develop risk assessments to address ethical issues linked to AI at all stages: inception, design, development, and, ultimately, implementation.
- Public awareness around the risks of use and non-use of data and AI are key to ensuring successful implementation of technological advances to drive achievement of the SDGs.
- It was noted that every AI needs data, including data for training AI models. While strong data protection and privacy frameworks need to be present and enforceable, a functioning system or mechanisms for making data available sustainably is one of the key steps to achieving AI for social good.
- Transparency on algorithms was demanded by many participants, especially with regard to the uses of AI by the public sector to inform decisions on citizens.
- Participants argued that there is a need to include a perspective of collective rights in Africa to complement the perspective on individual rights.
- Experts shared concerns about the expansion of digital IDs with biometric registration in different countries in Africa and the risks that might arise from a lack of policy and security safeguards and capacity.
- Participants emphasized the importance of engaging youth and investing in education programmes to build the technical capabilities of future generations, with a particular focus on education for women and girls.
We wish to thank Germany’s GIZ and Hewlett Foundation for funding this programme of work.