University of Ghana Vice-Chancellor urges inclusion of African languages in AI development

The Vice-Chancellor of the University of Ghana, Professor Nana Aba Appiah Amfo, has issued a critical call for the deliberate integration of African languages into artificial intelligence (AI) development. This imperative arises from the risk of Africa's marginalisation in the global AI revolution, which could lead to the underrepresentation of the continent's diverse voices and knowledge systems. This article explores the profound legal and ethical implications of linguistic exclusion in AI, examining how existing data protection, intellectual property, and human rights frameworks in Ghana and across Africa address, or fail to address, these challenges. It highlights the urgent need for robust regulatory interventions and proactive strategies to ensure equitable and inclusive AI development that respects linguistic diversity and prevents digital discrimination.

The rapid advancement of artificial intelligence (AI) presents both unprecedented opportunities and significant challenges for nations worldwide. In a recent address at the Fifth Warwick Distinguished Africa Lecture, Professor Nana Aba Appiah Amfo, Vice-Chancellor of the University of Ghana, underscored a critical concern for the African continent: the intentional inclusion of African languages in AI systems. Her warning that Africa risks marginalisation in the global AI revolution if its linguistic diversity and knowledge systems remain underrepresented resonates deeply within legal and policy circles, calling for urgent attention to the ethical and regulatory dimensions of AI development.

This call is not merely an academic plea but a strategic imperative with far-reaching legal, economic, and social implications. The absence of African languages in AI training data and applications can perpetuate biases, limit access to essential services, and undermine cultural heritage, thereby exacerbating existing digital divides. For legal professionals, Professor Appiah Amfo’s statement highlights a burgeoning area of law concerning digital rights, data governance, intellectual property, and the equitable application of emerging technologies. This article will delve into the existing legal frameworks in Ghana and the broader African context, analysing the gaps and opportunities for ensuring that AI development is inclusive, ethical, and aligned with the continent's developmental aspirations.

The legal landscape governing digital technologies in Ghana and Africa has been evolving, albeit often playing catch-up with technological advancements. Ghana's Data Protection Act, 2012 (Act 843), serves as the primary legislation for safeguarding personal data, regulating its collection, processing, use, and disclosure. This Act establishes a Data Protection Commission tasked with ensuring compliance with principles such as lawfulness, purpose limitation, data minimisation, and data subject rights, including the right to access and correct personal data. At a regional level, the African Union Convention on Cybersecurity and Personal Data Protection, commonly known as the Malabo Convention, adopted in 2014 and entering into force in June 2023, provides a harmonised framework for data protection and cybercrime across the continent. Ghana is one of the countries that has ratified this convention, indicating a commitment to regional data governance standards.

Beyond data protection, Ghana also has the Copyright Act, 2005 (Act 690), which protects original literary, artistic, and scientific works, and notably includes provisions for the protection of folklore as part of Ghana's cultural heritage. The rights to folklore are vested in the President on behalf of the people, requiring permission and fees for commercial use. This framework becomes particularly relevant when considering the use of traditional knowledge and linguistic expressions in AI development. Furthermore, Ghana has a National AI Strategy (2025-2035), which aims to position the country as an AI leader in Africa, emphasising inclusive growth, ethical AI, and alignment with the UNESCO Recommendation on the Ethics of Artificial Intelligence. The African Union has also adopted a Continental AI Strategy (July 2024), promoting an Africa-centric, ethical, and inclusive approach to AI, aligned with Agenda 2063 and the Sustainable Development Goals, and explicitly calling for the preservation of cultural heritage.

The call for linguistic inclusion in AI development highlights several critical legal challenges. Firstly, the exclusion of African languages from AI training datasets can lead to algorithmic bias and discrimination. AI systems trained predominantly on data from a few dominant languages may perform poorly or inaccurately when applied to African linguistic contexts, impacting access to justice, healthcare, education, and public services. The UNESCO Recommendation on the Ethics of Artificial Intelligence, adopted by Ghana, explicitly calls for fairness, non-discrimination, diversity, and inclusiveness throughout the AI system lifecycle, urging efforts to minimise and avoid perpetuating discriminatory outcomes. This international instrument provides a strong ethical and normative basis for advocating for linguistic diversity in AI.

Secondly, the collection and processing of linguistic data, particularly indigenous languages and oral traditions, raise complex data governance and intellectual property issues. While Ghana's Data Protection Act, 2012 (Act 843), provides a framework for personal data, the specific nuances of linguistic data, especially when it embodies collective cultural heritage, may not be fully covered. The Act's principles of consent and purpose limitation are crucial, but questions arise regarding collective consent for community-owned linguistic assets and the appropriate attribution and benefit-sharing mechanisms. The Copyright Act, 2005 (Act 690), offers some protection for folklore, vesting rights in the President for the people of Ghana. However, adapting these provisions to the context of AI training, where vast amounts of data are ingested and transformed, requires careful consideration to prevent exploitation and ensure equitable benefit-sharing with linguistic communities.

Thirdly, the broader human rights implications of linguistic exclusion in AI are significant. The African Charter on Human and Peoples' Rights, which Ghana is a signatory to, guarantees the right to economic, social, and cultural development, rooted in freedom and identity. Digital marginalisation due to linguistic barriers in AI systems could be argued to impede these rights. The African Commission on Human and Peoples' Rights' Declaration of Principles on Freedom of Expression and Access to Information in Africa (2019) further states that the development and use of AI should be compatible with international human rights law and standards. This implies a legal obligation for states to ensure that AI systems do not infringe upon fundamental rights, including those related to language and culture.

Comparative approaches, such as those being explored by ECOWAS, indicate a regional understanding of these challenges. The ECOWAS Parliament is actively pushing for a community-wide legal framework for AI, including a "West African Pact on Ethical AI and Digital Education," with a focus on promoting AI tools in local West African languages. This regional initiative, alongside Ghana's National AI Strategy and the AU Continental AI Strategy, signals a growing recognition of the need for tailored regulatory responses that address the unique linguistic and cultural contexts of Africa. However, the challenge lies in translating these strategic aspirations into concrete, enforceable legal instruments that mandate linguistic inclusion, ensure data sovereignty over linguistic assets, and establish clear accountability mechanisms for AI developers and deployers.

Professor Nana Aba Appiah Amfo's call for the inclusion of African languages in AI development is a timely and critical intervention, highlighting a complex interplay of technological advancement, cultural preservation, and fundamental rights. For legal practitioners, this presents a burgeoning field requiring expertise in data protection, intellectual property, human rights, and emerging AI governance. Attorneys must be prepared to advise clients on ethical AI development, ensuring that data acquisition for AI training respects linguistic diversity, obtains appropriate consent, and adheres to principles of fairness and non-discrimination as enshrined in national laws like Ghana's Data Protection Act, 2012 (Act 843), and international instruments like the UNESCO Recommendation on the Ethics of AI.

Looking ahead, the legal landscape will likely see increased efforts to develop specific AI legislation that addresses linguistic inclusion, algorithmic bias, and cultural heritage protection. Practitioners should monitor the implementation of Ghana's National AI Strategy and the AU Continental AI Strategy, as well as ECOWAS initiatives, for new regulations and guidelines. There is a clear need for legal innovation to craft frameworks that not only mitigate risks but also actively promote the development of AI systems that are truly representative of Africa's rich linguistic and cultural tapestry, ensuring that the continent is an active participant, rather than a passive recipient, in the global AI revolution.