During this dynamic time of COVID-19, it is concerning how some African countries do not have any established Artificial Intelligence(AI) policies towards AI implementation, evaluation or monitoring. The main question is how they are planning to move forward especially when AI solutions are being proposed to mitigate various needs on the ground. Are African governments and their respective support structures in touch with the whole process of AI policy formulation or are they facing the dilemma of implementing AI solutions without having appropriate policies in place?
The question for policymakers that we are posing today is a process question. Are there any standard processes for AI policy formulation and implementation during critical times? What is the way forward when no such policies exist in the first place? If we take a step back and assess the whole cycle of technology development, technology implementation is often always leading while policy formulation would subsequently follow. This mismatch between the pace of technology advancement and the policy processes are a cause for concern for many technology developers and investors alike.
In the case of AI, if we handle it with a socio-technical arm, which entails the consideration of complex social structures, human values, norms or roles to inform the technical design of AI driven solutions, it will be apparent from the policy perspective that the whole process should be consultative in nature and all concerned stakeholders, be they policymakers, end users or developers, need to have a say in decision making. This brings to the fore an important aspect of ethical issues around AI and how policy formulations should address them.
Virginia Dignum’s work on Responsible AI includes recommendations for policymakers on how to address AI policies and values from inception onwards. The key takeaway from Virginia Dignum’s work is how we can ensure that policies are not always reactive when it comes to socio-economic issues and more importantly when we add a socio-technical-economic dimension to the equation. At times like these, there is an urgent need for responsible AI solutions yet without appropriate policies in place, the problem of violating ethical aspects of AI could later arise. In a collaborative and consultative atmosphere, all concerned parties can equally contribute to timely and comprehensive solutions.
Recommendations for Policymakers
Despite the common belief that the policymaking process is time-consuming, it is important that policymakers get involved from the onset to pave the way for innovative solutions. It is high time for policymakers and technology developers to work together towards a common goal. Let this not be an initiative exclusively at times of crises, but a best practice for all actors to formulate sustainable, responsible and accountable solutions enabled by various frontier technologies.
About the author
Dr Attlee Gamundani is a Young ICTD Fellow at the United Nations University Institute in Macau. His research interests revolve around the issues of Artificial Intelligence of Things (AIoT), Cybersecurity, ICT for development and Sustainable Development Goals (SDGs).