Moonshot: Africa should not play catchup with AI regulation

Moonshot by TechCabal is the conference that brings together Africa’s tech ecosystem in person to network, collaborate, share insights and celebrate innovation. Join us in Lagos on October 11 and 12. In this third article built around the conference, Kenn Abuya considers how African governments can adopt a united approach with AI regulation while upholding human rights, ensuring data privacy and protecting individuals’ intellectual property.

Since the early days of the internet, Africa has mostly been a consumer of new tech. Like in social media, where non-US nations have to adapt to American platform rules, and Africa often has little say in shaping or challenging these norms, the same pattern is happening with powerful AI systems like ChatGPT. Africa is again in the receiving position. But, this time, the continent can shape the regulations to position itself as an active and influential participant in the global AI landscape. This involvement can lead to a more equitable, ethical, and responsible AI development and deployment across the continent. Africa can push for a new and inclusive approach to regulate AI, unlike how it tackled social media.

AI regulation in Africa has a long way to go

African nations, such as Kenya, have yet to address the regulatory aspects of AI. While some existing legal frameworks offer potential avenues for incorporating AI elements, the rapid advancement of this technology has outpaced the scope of these laws, particularly concerning data collection. Given this scenario, creating a dedicated, all-encompassing regulatory framework for AI is key. Kenya already has some regulations in place, such as the Data Protection Act (2021) and the Computer Misuse Act (2018), which could be amended to accommodate the dynamics of AI applications.

According to J. Walubengo, ICT lecturer at Multimedia University of Kenya, who spoke to TechCabal, “AI is quite a broad space. A new, isolated Act to regulate AI would be needed.” Walubengo cites the proposed EU AI Act as a framework that can guide the development of AI regulations for the African market. For instance, its goals include ensuring safety and adherence to existing laws regarding fundamental rights and values for AI systems entering the EU market. The framework also seeks to establish legal clarity, promoting an environment conducive to investment and innovation in AI. It also aims to bolster governance and enforce prevailing laws concerning fundamental rights and safety prerequisites applicable to AI systems.

For now, a few African states have made significant progress regulating AI: Tunisia, Egypt and Mauritius. While the call for Africa to engage in AI regulation is valid, Africa must also consider the practical challenges of the continent’s digital policy decisions and the lack of necessary infrastructure that affects the development of AI legal frameworks, Megan Kathure, a tech policy analyst, argues.

“AI regulatory attention by African states will take various forms, as already evidenced by Mauritius and Egypt, which have National AI strategies, while other countries adopt a piecemeal approach to governance by creating task forces or research labs. While the clamour for the continent to throw its hat into the ring of AI regulation bears merit, sight should not be lost on the realpolitik attendant to Africa’s digital policy-making and the infrastructure deficit impacting the coming into fruition of AI legal frameworks,” Kathure told TechCabal.

Most African nations, like Kenya, have not yet implemented a formal national strategy. However, Kenya has introduced recent resources like a guide for AI practitioners. It’s important to note the significance of the Blockchain & Artificial Intelligence taskforce discussed in 2018, which wrote a report called “Emerging Digital Technologies for Kenya: Exploration and Analysis.” It offers guidance for individuals engaged in AI development and use. It also aids in shaping future regulatory initiatives by Kenyan authorities.

Ethical AI regulations are important, though

The nature of AI systems poses a challenge to lawmakers seeking to establish regulations that safeguard individuals’ rights. However, one key question emerges: how might legislators navigate the nuances of AI and applications to create regulations that uphold human rights?

According to Walubengo, specialised legislation often overwhelms legislators. The solution involves the legislators forming a task force of AI experts, legal professionals, and academics to draft the law. They would then educate the parliament to enable informed debate and passage of the legislation. Megan Kathure, on the other hand, argues that effective lawmaking in this era of emerging technologies should consider how these technologies impact various aspects of human rights. Recognising this and prioritising human rights in policymaking will assist policymakers in navigating AI regulation.

“Legislators usually can’t cope with most specialised legislation; what happens is that the minister for ICT would need to constitute a taskforce of AI experts/practitioners, legal experts, AI academics, etc, to draft the law, then educate the parliament on the same for them to debate and pass the legislation,” said Walubengo.

“Effective lawmaking in this epoch of emerging technologies should take note of the rapid diffusion of technologies in every facet of our lives and how such incursion has the potential to facilitate or stymie the realisation of human rights. Such an appreciation, coupled with the intentional centralisation and respect for human rights in policymaking, will aid policymakers in navigating the spheres of AI regulation,” added Kathure.

AI can be impactful if properly regulated

AI can wield significant influence when subject to effective regulation. Proper oversight ensures its potential benefits are harnessed while minimising risks, fostering innovation, and safeguarding ethical considerations. Walubengo acknowledges AI’s value for productivity, but with a caveat: its success hinges on extensive data, and some AI firms have already breached privacy laws to gather vast personal and potentially copyrighted data. Therefore, regulation is vital to harmonise the interests of AI developers, personal data, and intellectual property.

He said, “AI is good and should be embraced and encouraged. However, its side effects (e.g., non-responsible AI) should be regulated for the benefit and safety of humanity.”

“While too often we wax lyrical of how regulation can stifle innovation—a plausible reality but often a stand-in for entities to have free reign—any sound AI regulation at its crux should act as drivers of innovation and be rights-respecting,” Kathure added.

It is all about balancing data privacy and intellectual property

Embracing the power of AI requires a nuanced approach. As we venture into this digital frontier, ensuring equilibrium between data privacy and intellectual property rights becomes key. Regulations must serve as the guiding force, enabling innovation to grow while safeguarding personal information. Striking this balance allows AI to thrive ethically and effectively, fostering an environment where technological advancement respects individual autonomy.

“Much of the interplay between data privacy, intellectual property protection and AI will depend on regulators’ agility to enforce data protection and intellectual property laws. This, again, will be affected by the ability of the same regulators to stave off corporate and geopolitical influence as well as their resourcing by national governments. There have also been repeated calls for African legal regimes to transpose African values into data protection and integrate the ubuntu philosophy into AI ethics, it would thus be worthwhile for such value transpositions to be effected. As I see it, it is feasible but will take time to materialise,” Kathure clarified.

Walubengo mentions the vast amounts of data needed to train AI and the need to enforce privacy laws while respecting intellectual property. “Indeed, it will help increase productivity. However, it requires massive amounts of data as raw material. Many AI companies violated data privacy laws to acquire massive amounts of data, including personal data. Additionally, some of the data harvested may include intellectually protected data sets. Hence the need to balance out the different stakeholders (AI innovators, citizens’ data, and intellectual property holders) interests through regulations,” he added.

Africa should unite its approach to comprehensive regulation and advocate for a collaborative global protocol shaped by its active involvement. Safeguarding intellectual property such as trademarks and copyrights is key, and by joining forces, Africa can see AI’s benefits unfold while safeguarding its unique interests.

Did you enjoy this article? Then click this link to register for Moonshot and check out our fast-growing list of speakers coming to the conference!