Topics covered during the event included the role of AI in social science and health research, the benefits and challenges of using AI and data science, and related ethics and governance issues in Africa.
The event’s objectives included how social science research could promote responsible use and adoption of AI, particularly in terms of gender and diversity, education and socio-economic development, understanding AI and data science gaps in global health from an African perspective and identifying potential threats and opportunities in the field.
Maiyane, the keynote speaker, is an expert in the role of AI in social science research. His doctoral thesis ‘Robots and Dignity from an Afro-Communitarian Perspective’, investigates the challenges around the development and use of AI technologies, specifically in terms of how it affects the dignity of people who use them, or for whom they are used.
COVID-19: a lesson for the future
The global pandemic was an exercise in using a scientific approach to resolving a critical health crisis, he said, and threw into focus the benefits and challenges presented by AI and the use of data.
The haste with which the global community came together to intervene and to meet challenges was commendable, with quick-step solutions implemented with speed: lockdowns, quarantines and intense study of the virus itself – a process that was inextricably attached to the process of vaccine development.
Maiyane said that the data tracking systems emerging from the event showed how useful technology could be: data showed how and where the pandemic moved, enabling stakeholders to develop intervention systems.
For example, he explained, the Department of Health and The Council for Scientific and Industrial Research (CSIR) launched a data centre to upload and project infection rates in real time.
With large numbers of health care professionals, experts and data scientists working together, a post-mortem revealed a global lack of preparedness in dealing with an event of this magnitude, he said.
Key takeaways from COVID-19 could inform how we approached AI and the use of data, explained Maiyane.
Some crucial solutions included technological assistance in tracking, diagnosing, predicting and preparing for large-scale events and the use of data and analytics for intervention and diagnostic purposes.
AI: watchdog needed
Collection and collation of data was all well and good, said Maiyane, but the use of AI gave the world as many challenges as opportunities.
Firstly, the harnessing of existing data to develop diagnostic and predictive models was laudable, particularly with cloud-based services ensuring data accessibility in real time, but the tricky ethical issues such as privacy, security, bias and opacity had to be addressed.
“How do we ensure that data is cultivated and shared so that different people, in different parts of the world, can use it for research and clinical trials? And how do we protect information in specific countries?
“There are also contextual challenges: bias towards specific groups in society, based on demographics such as race, class and gender, and consequences emanating from that; people’s risk profiles are determined based on the colour of their skin, gender or where they live.
“Because we don’t produce most of this data ourselves, we cannot determine how certain conclusions are reached based on data gathered and interpreted elsewhere.”
The role of social science
Social scientists ask different questions to those asked by others, said Maiyane, and these questions were critical if AI was to be utilised fairly and beneficially.
Questions included: who is affected, who benefits, how does this affect societal well-being, and what does data expose about us?
In terms of health, an African social science priority was to find effective health care solutions that specifically addressed inequalities on a continent where poor people could not access all aspects of modern health care, he said.
Social science research could, in addition, assist governments in navigating ethical issues such as privacy and bias, optimise user experience and inform reviews for updating policies and regulations.
The launch of The Presidential Commission on the Fourth Industrial Revolution (PC4IR) was an excellent starting point, but more investment in social science research on AI, and a rigorous review of old policies and regulations, plus new ones, were also key, he said.
Social scientists offered context-based solutions, rather than empirical data based ones, and this was valuable in the AI space, he said. “We look at the impact of the product on the user, and not just the product itself.”
A human-centred approach to AI – literally putting a human face on it – was the best way forward.
Other speakers at the event were Science for Africa Foundation’s Dr Uzma Alam, who spoke on policy status and gaps in AI and data science in Africa, and the HSRC’s Dr Mokhantso Makoae, who spoke about the role of gender and diversity in AI and data science in both social science and health research.
About Karabo Maiyane
Karabo Maiyane is a lecturer in philosophy at Nelson Mandela University. His teaching expertise includes moral and political philosophy as well as the history of philosophy. He is a doctoral candidate at the University of Pretoria. His research focus is on the ethics of artificial intelligence (AI). He is interested in AI technologies' impact on human dignity in the context of healthcare, warfare and companionship. His other interests are ethics of technology, just war theory, ethics of care, healthcare ethics and African ethics in general. He has recently published two chapter contributions titled: "Robots and Dignity: An Afro-communitarian Argument in Eldercare" and "Autonomous Weapons and the Future of Warfare in Africa".