University of Pretoria Focuses on Responsible AI Implementation Through Transdisciplinary Research
The University of Pretoria is concerned about balancing innovation and accountability to ensure the responsible use of artificial intelligence (AI)
This is a theme that is not just of interest to the University’s computer science disciplines in the Faculty of Engineering, Built Environment and Information Technology (EBIT). The legal implications related to the responsible use of AI are of interest to UP’s Law Faculty as well.
Joining forces to develop a two-pronged approach to tackle this theme are Prof Vukosi Marivate, holder of the Absa UP Chair of Data Science in EBIT’s School for Information Technology, and Dr Chijioke Okorie, leader of UP’s Data Science Law Lab. These two researchers believe that AI producers need to better consider the communities that directly and indirectly provide the data used in AI development.
Their recent focus, which resulted in an article published by Carnegie Endowment for International Peace, examined how African natural language processing (NLP) experts are navigating the challenges of copyright, innovation and access.
In the introduction to their article, they explain how, as an ideal or a practice, openness in AI involves sharing, transparency, reusability and extensibility that can enable third parties to access, use and reuse data, and deploy and build upon existing AI models. This includes access to developed datasets and AI models for purposes of auditing and oversight, which can help establish trust and accountability in AI when it is done well.
Prof Marivate explains how some common sayings in African languages encapsulate how issues of agency and community ownership are implicated or threatened when openness is embraced in a bid to include Africa and other parts of the Global South in discussions about the responsible use and development of AI.
While actors in the Global North have been the primary drivers of discussions about responsible AI by focusing on concepts like openness, privacy and the protection of copyright, there have been increased efforts to amplify perspectives from under-represented jurisdictions. This is in an effort for them to help shape discussions about responsible AI use and development in the Global South. Within this atmosphere of inclusion, openness, privacy and copyright have continued to feature as important and indispensable considerations. This is referred to as the Global South inclusion project.
To highlight these interests and concerns, the research featured a study of the African NLP community, presenting insights from the work of the Masakhane Research Foundation, a distributed research organisation with the mission to advance African NLP, as well as Ghana NLP, an open-source initiative focused on NLP involving Ghanaian languages, and KenCorpus, a community-driven project to create large Kenyan language datasets. The experiences of this community would help ground the practical trade-offs and challenges that arise in this discipline.
The research furthermore highlighted some of the opportunities and challenges presented by considerations around openness as a way to address copyright and privacy concerns. Openness must be practiced in a manner that considers the communities that provide the data used in commercial and non-commercial settings for AI development. Depending on the use case, the interests of these communities may involve financial and social benefits, or mere attribution or acknowledgement.
As a result of their proprietary and rule-based nature, copyright and privacy rules may result in practices that discourage openness. Yet, addressing the restrictive and proprietary nature of these rules through openness does not mean that openness is adopted without attending to the nuances of specific concerns, contexts and people. In adapting openness to the nuances of the contexts of Africa, consideration must be given to the agency and autonomy of specific stakeholders to make decisions about the uses of their data contributions, created and annotated datasets, and the needs that AI tools and development are designed to address. The intersectionality of these concerns necessitates a comprehensive approach to data governance, which addresses the multifaceted challenges and opportunities presented by Africa’s evolving data landscape.
From a regulatory point of view, copyright and privacy laws may need internal reforms, explains Dr Okorie. There may even be the need for a specific piece of legislation, such as the European Union has undertaken with its recently passed Artificial Intelligence Act. However, of more immediate benefit, given the protracted nature of legislative reforms, is the use of contracts and private ordering regimes. “The doctrine of freedom of contract means that changes and tweaks can be made in existing open licensing regimes to address relevant challenges and harness relevant opportunities.”
The researchers conclude that the good news for private actors is that they can directly make changes and tweaks in the open licensing regimes to address the challenges and harness the opportunities encountered in this research.