Need for fundamental regulation for the Global South
1 Dec 2022 08:15h - 09:45h
The meeting aimed to discuss ways of empowering the Global South to create inclusive AI technologies that serve the public interest. It examines the opportunities, risks, and use cases of digital public goods (DPG) for AI and, on the policy front, the conceptual and normative frameworks that build and curate spaces for discussions around AI regulations. It concludes on ways to guarantee DPGs institutional capacities.
On the socioeconomic front, in the Global South, gains from the use of AI technologies are evident, particularly in traditional areas like finance and banking, where regulations are generally imposed at the global level. In Ethiopia, for example, AI technologies are deployed with the purpose of easing the burdens associated with difficult tasks. Different from the technologies constructed in the global north, these homegrown technologies are designed to cater to the 92 languages spoken in Ethiopia. The adoption of AI technologies can also be seen in other fields such as agriculture, education, health, and tourism, and have also registered marked returns, such a case being the Brazilian Judicial system.
Despite the ubiquitous enthusiasm with present and anticipated future gains, many are anxious about the known and unknown negative risks associated with the implementation of AI technologies by government, corporations, and private citizens. The anxiety and mistrust surrounds, in the first instance, the blackbox nature of AI technologies and the hesitancy demonstrated in developed countries like the USA. The fact that the majority of AI technologies were trained on data and experimented upon in ecologies in the global north is also cause for concern, given its capitalist motivations. Currently, much of the anxiety is centred on tangible technologies and focuses on issues of privacy and surveillance. Questions arise concerning the points of who holds the data, for what purpose(s), and for how long as well as whether private data is being used to replace humans in the job market or for the benefit of corporations and political power.
Concerning the regulations for AI technologies, there is widespread consensus that outside of the traditional sectors previously referred to, regulations are at worst non-existent or ad hoc at best. The meeting agreed that not only is there a need to formulate national regulations/policies to address AI usability and utility in all sectors, but there is a need to approach policymaking from a systems perspective as opposed to the prevailing technical lens currently being used to frame existing AI principles.
The framing of regulations must consider the full gamut of the technology use case, including cultural and social norms and needs, without stifling innovation and stymying the economic and social gains that come with AI. Such requirements render the framing of AI regulations complex, costly, and time consuming, and would demand transparency and accountability at every level. For the global south, the concern here is whether those currently being trained in AI technologies are being exposed to the ethical and explanatory issues surrounding AI product adoption, and whether the human in the loop, at all points of the value chain, could be effectively trained so as to mitigate or prevent the perceived harms associated with AI products. Overall, in order to decrease the widening gap between policy and innovation and public trust and the deployment of AI technology products, an open approach to government arrangements is needed, corporations must infuse ethical and culturally sensitive principles upstream in AI technologies products, and a multistakeholder approach is required in regulation formulation, implementation, monitoring, and evaluation.
By Alicia Shepherd
The session in keywords
Internet Governance Forum 2022
28 Nov 2022 - 2 Dec 2022
Addis Ababa, Ethiopia and online