From telemedicine to advanced diagnostics, artificial intelligence (AI) is reshaping the healthcare landscape across the globe. In African nations, there is a burgeoning interest in the use of AI for health research and patient care. Yet, this technological leap is met with regulatory dilemmas. A recent study by my research group, involving twelve African countries, reveals that while all are in the early stages of their AI regulatory journey, there is significant variation in their progress.

A patchwork of regulations

One major finding is the diversity in AI regulatory frameworks across these nations. While many have made strides in areas like data protection, crucial for AI development, there’s still a fragmented landscape. This poses challenges for companies and institutions aiming to collaborate across borders.

Drawing inspiration from global players

While AI-specific regulations may still be in their infancy in Africa, other global players offer valuable insights. The EU, for instance, has been proactive with its proposed ‘Regulation laying down harmonisation rules on Artificial Intelligence’ and the US with its ‘Algorithmic Accountability Act’. African nations can draw inspiration from other countries, while adapting to their unique challenges.

Who owns AI-generated innovations?

The rise of AI in healthcare promises innovations, from drug discoveries to diagnostic tools. But an age-old question gets a futuristic twist: When AI creates, who owns the rights? The current intellectual property (IP) laws in many countries, including those in Africa, base rights on human creation. AI-generated outputs muddy these waters, leading to debates about ownership and rights. The fact that South Africa is the first country in the world to have awarded a patent to an invention where the inventor is an AI system — and not a human! — is noteworthy. However, policies need to be developed to clarify this issue and promote innovation in Africa.

African Voices Matter

As AI evolves, it is crucial to avoid perpetuating biases. The AI we deploy in healthcare should be free from prejudices that might inadvertently deepen societal divides. The need to address algorithmic bias is paramount. The data that feeds AI can either uplift or marginalise communities. It is a challenge that’s both technical and moral.

Moreover, it’s essential to recognise that while AI can be transformative, its design and application should be respectful of Africa’s diverse cultures and histories. Embracing multiculturalism in AI governance ensures that the technology serves all communities equitably.

Conclusion

The road to integrating AI in African health research and healthcare is filled with both promise and pitfalls. With a balanced regulatory approach, a commitment to inclusivity, and an eye on global standards, Africa stands poised to harness the power of AI in healthcare, charting a course for the rest of the world to follow. The future beckons, and it is one where technology and humanity walk hand in hand.

This blog post is based on an in-depth research study by my research group that delves into the nuances of AI regulations in twelve selected African countries.

A decade ago, the promulgation of the Regulations relating to the Artificial Fertilisation of Persons marked a significant moment in South African reproductive law history. Regulation 18, in particular, introduces a scheme of ownership in reproductive material, such as sperm, eggs, and embryos. But can this ownership, once acquired, be transferred? I explore the complexities surrounding this question in this article in our country’s premiere law journal — the South African Law Journal. Here is a summary:

Ownership, in legal terms, is a bundle of rights encompassing the power to use, enjoy, consume, possess, dispose, reclaim, and resist any unlawful invasion of the object. But how is ownership acquired? South African law divides the acquisition of ownership into original and derivative forms. Derivative acquisition, where ownership is transferred from one party to another, is common in daily transactions. Original acquisition, however, occurs when a new legal object comes into existence. Consider the birth of a puppy. The puppy, before its birth, exists physically but not legally. Only upon its birth does it become a legal object, and ownership is vested in the mother dog’s owner. This illuminates a fundamental legal notion where objects must have independent existence to be owned legally. Similarly, in the context of reproductive material, ownership begins when gametes are separated or an embryo is created.

Regulation 18 provides who acquires ownership in the gametes and embryos. But, does this regulation allow for the transfer of ownership by one person to another? There are two interpretations of regulation 18: the supple interpretation, allowing the default allocation of ownership to be changed, and the rigid interpretation, allowing no change to the default position. If we consider the rigid interpretation, we encounter complications that pose serious legal challenges. Scenarios such as abandonment or unexpected death present conundrums that demand legal flexibility. These practical problems push us toward the supple interpretation. Furthermore, the supple interpretation enables freedom of contract and promotes values of freedom and dignity, aligning with South Africa’s human rights regime. The rigid interpretation, in contrast, seems anathema to these values, suppressing freedom and promoting no discernible constitutional value.

In conclusion, the issue of transferring ownership in reproductive material is a legal labyrinth. The supple interpretation appears to be the path forward, allowing for flexibility and alignment with constitutional values. Reproductive law, full of potential and ethical intrigue, beckons further exploration and invites us to reflect on the intricate relationship between law, ethics, and humanity.

The burgeoning landscape of research, particularly in genomics, calls for the careful balancing of the rights of data subjects with scientific imperatives. South Africa is rising to the challenge, with the development of a Code of Conduct for Research in terms of the Protection of Personal Information Act 4 of 2013 (POPIA) — currently in draft form.

The draft Code of Conduct for Research is a welcome development, but there is room for improvement. My research group identified three broad areas where improvement is needed, considered potential solutions, and authored an article on each of these areas: First, the way that the draft Code describes core concepts in POPIA requires refinement. Second, in line with the global trend towards open science, there are significant benefits to an open access genomics database of South Africans. The draft Code should provide a clear roadmap for open access genomics projects. By doing so, it will not only foster innovation but also align the South African research environment with global best practices. Third, the need to strengthen the Code of Conduct also emerges from the pivotal role that research should play in the South African economy. Research is not (merely) a purpose in itself, but a fundamental contributor to the South African economy. Because research data — which may consist of personal information — often has commercial value, it is important that we consider legal pathways for repurposing research data for commercial use. The draft Code should provide guidance in this regard.

In conclusion, the draft Code marks an important milestone but needs further refinement to reach its potential as a practical tool for researchers. The revised Code could set a robust framework that safeguards privacy, promotes innovation, and contributes to the economic vitality of South Africa.

The issue of ‘benefit sharing’ with research participants is mired is controversy. In conventional ethics, donating biological specimens and personal data to scientific research is viewed as a selfless act aimed at promoting the greater good. However, proponents of ‘benefit sharing’ argue that scientific research is a lucrative business and that considerations of distributive justice demand that research participants who donate their biological specimens and personal data to scientific research projects should receive ‘benefits’ in return. But would such ‘benefit sharing’ not risk turning research participation into a commercial activity that undermines the dignity of the human body? And would the promise of receiving a ‘benefit’ not compromise individuals’ autonomy to decide based on their own authentic values whether to participate in research? Accordingly, the social harm of practicing ‘benefit sharing’ may be most harshly felt in poverty-stricken communities.

There are also economic policy arguments that are relevant: While international pharmaceutical companies certainly make fortunes from their scientific research, the same is not true of local universities and biotech startups. As such, while international pharmaceutical companies may easily be able to afford giving all kinds of ‘benefits’ to their research participants, this is likely to create expectations among research participants that local universities and biotech startups would not be able to compete with. This will put a damper on South Africa’s aspirations to build a local bio-economy. In the long term, building a vibrant and sustainable bio-economy is key to South Africa’s participation in the Fourth Industrial Revolution and lifting our people out of poverty.

Importantly, South African law presently outlaws ‘benefit sharing’ with research participants in research projects that entail donation of biological specimens. This decision by South Africa’s democratically elected parliament should be respected by all involved in research in South Africa.

The idea of genomic sovereignty entails that a nation state should control the genomic resources found in its territory as a national asset. This is established international law in the context of non-human genomic resources. However, some have argued that it should also be applied to human biological samples and genomic data. This raises some serious human rights concerns. In Mexico, the idea of genomic sovereignty was indeed applied in public policy on human genomic data. In our recent article, Dr Faith Kabata and I compare Mexico’s public policy experiment with a different approach, namely that of Finland. In contrast with the state-centred nature of genomic sovereignty, the approach adopted by Finland embraces private ownership of genomic data by local biobanks, subject to control of such data by the individual data subjects themselves. We suggest that this approach is more promising than genomic sovereignty, as it better aligns with the rights of individuals — in particular their right to participate in and benefit from scientific progress and its applications.