News Perspectives on the Future of Responsible AI in Africa

RAIN-Africa participated in an event co-hosted with the Montreal AI Ethics Institute (MAIEI) on December 14th. The Perspectives on the Future of Responsible AI in Africa event was structured as an diverse and extended discussion that gave insights into how responsible AI could be promoted in Africa and how understandings of ethics and social impact might have to be revised when it comes to the implementation of AI-enabled tools in Africa.

After a short introduction by Connor Wright (MAIEI), attendees were split into two breakout rooms to discuss and develop solutions to the following questions:

  1. Given the opportunities and challenges for developing and using AI in, what should be done NOW to promote responsible AI?
  2. Does our understanding of Responsible or Ethical AI need to be fundamentally adapted to apply to the social concepts and historical biases that underlay many Africa cultures, such as Ubuntu? Where are their overlaps with international ethical concepts and where do we need to adapt?
  3. How do we build strategic partnerships to understand the responsible use of AI for sustainable development?

The attendee were split into two groups to promote more in depth discussion. Each had  representatives from RAIN-Africa, with Prof. Christoph Lütge and Laud Ammah in one room and Prof. Jerry John Kponyo and Dr. Caitlin Corringen in the other.

After 45 minutes of discussion, attendees were brought back together to discuss important findings from their groups and summarize the event.

Some of the major points raised during the discussions included:

  •  The lack of datasets derived in Africa (used, for instance, to benchmark facial recognition technology ).
  • AI is being driven by the US/western developers and companies. Therefore, there is a need to tackle this challenge first before discussing further regionalization within the African continent.
  • Many of the well-known ethical frameworks do not take things like „community” into account and instead focus more on individual impacts and rights. Since ethical considerations with AI are often about tradeoffs, developers and users of AI in Africa might have to consider how their ethical priorities match those of the often used ethical AI guidelines.
  • The impact of AI will be minimal if it is not attuned to the culture/context it is being used in. There is a need to understand the environment in order for the AI application to be human centered. This involve interdisciplinary approaches to AI development that rely not only on technical, but also cultural expertise.
  • We cannot avoid the oncoming AI future, especially when it comes to incorporating it into higher education curriculums. We should care more about teaching students and workers about the potentials AND risks of employing AI-enabled tools.
  • There is also a lot of work needed to be done on the government’s side. We cannot expect people developing ethical solutions if we are involved in an inadequate environment, such as corruption.
  • There is a lack of understanding and current institutional weakness needs to be remedied to promote responsible AI.