UNESCO Working Session: Priority Africa Flagship Programme

AI risks for Human Rights, Ethics of AI, EVENTS, Priority Africa, Uncategorized

Professor Claudia Roda gave a short presentation at UNESCO’s Working Session Priority Africa Flagship Programme 4: Harnessing new and emerging technologies for sustainable development in Africa, including through the implementation of the Recommendation on the Ethics of Artificial Intelligence.

A few points made during the talk

Promising technology developments in AI:

  • Low(er) resource and open access models start democratizing access to generative AI
  • Specialized systems using better (selected) data start producing more reliable results
  • Local (off the cloud) data and computation would allow better data protection and security

AI has a great potential for improvements in areas such as scientific research, healthcare, and agriculture, but also enormous risks that we need to address before the deployment of AI systems. Risks include widening the digital divide, exacerbating economic differences, increasing geopolitical asymmetries, worsening the environmental crisis, and the concentration of strategic digital resources and services among a small number of individuals, businesses or states.

I am here to listen and learn, but I want to share some of the guiding principles for our Chair which also apply well to the strategy for Africa:

  • We support the implementation of the UNESCO recommendation on the Ethics of AI but also a set of other instruments already available and being developed by UNESCO addressing issues of governance and risks (bias, gender, transparency)
  • AI is always a mix of AI and non-AI technology; we need to assess both
  • When we assess we need to pay attention to software but also to hardware (e.g., advantages/disadvantages of different infrastructures) – in Africa, the issue of mineral extraction raises important questions: environmental, labor, exploitation, …
  • We must consider all solutions, including low tech or no tech: evaluate advantages and risks (e.g., The work of the Chair is supported by Lifeline energy, organization focusing on low tech services providing solar-powered radio in Africa)
  • AI (like all technology) is not location/culture/language agnostic, when we design, implement, deploy, maintain, assess AI systems we need to pay attention to local needs