On 25 June 2025, Unisa’s Directorate Language Services (DLS) hosted an Artificial Intelligence (AI) Language Symposium exploring the theme Multilingual Language Practice and Artificial Intelligence: Lost in Translation or Found in Meaning?

The purpose of the symposium was to gather AI experts, language scholars, researchers, academics and language professionals to discuss and share knowledge on the impacts of AI on language teaching, learning and research in institutions of higher learning, specifically regarding the indigenous languages, and other related topics. It was a platform for exchanging ideas, exploring new perspectives and fostering collaboration, particularly in an open and distance e-learning (ODeL) environment.

Prof Meahabo Magano, Executive Director of the Department of Tuition Support and Facilitation of Learning (DTSFL), delivered the opening and welcoming remarks. She expressed her excitement about the strides taken by Unisa in embracing multilingualism. Her demonstration of a "Student Walk" emphasised the role of multilingualism in higher education in taking students through their journey for epistemic access.

Prof Meahabo Magano

Magano highlighted the need for decolonisation of the mind in epistemology and to advance the use of language technologies and AI in teaching, learning and research. She further referenced Ngugi Wa Thiong’o’s concept of "decolonisation of the mind", which emphasises the importance of language in the process of liberating oneself from the mental and cultural shackles imposed by colonialism, arguing that African writers and scholars’ continued use of European languages perpetuates a "neo-colonial slavish and cringing spirit".

Dr Rakwena Monareng, Director of Language Services, presented a brief overview in which he unpacked the theme of the AI Language Symposium, the core ideology and the mandate of DLS. He reiterated the need for AI in a digital world and the fact that Unisa, as a multilingual institution, cannot be monolithic in approach, hence multilingualism and the multilingual language policy implementation. Monareng stated that the indigenous South African official languages should be central to knowledge making and sense making. He advised that AI, as a transformative tool, should be infused in teaching, learning and research.

Dr Idris Abdulmumin delivered the keynote address on behalf of Prof Vukosi Marivate from the University of Pretoria (UP). The topic was How we promote multilingualism and infuse AI. Abdulmumin showcased projects initiated at UP in pursuit of integrating AI applications in the digitisation and digitalisation of indigenous languages. He further premised the reasons why African languages are characterised as low-resourced, lacking data sets, having weak language policies, and lacking computing power for developing natural language processing (NLP) models. His presentation's highlight was the decolonising science project, whereby science terminology was being translated into six indigenous African languages to allow full automation and availability of the referenced languages on online AI platforms.

The Director of the South African Centre for Digital Language Resources (SADiLaR), Prof Langa Khumalo, was one of the respondents after the keynote address. His presentation entitled African Languages in Education in the Context of the Digital Revolution emphasised what a language is and how best it can be utilised as a tool for information dissemination in a multilingual environment. He explained that SADiLaR supports creating, managing and disseminating digital language resources and applicable software in all official South African languages. Khumalo emphasised multilingualism as a resource that affirms every student in higher education institutions, and that linguists and language professionals should take the lead in developing language resources that can be easily accessible to all students. He reiterated the significance of inter-institutional collaborative programmes and the sharing of language resources and terminologies. He gave the example of a SADiLaR-Wikipedia-PanSALB (SWiP) project to promote resource-scarce languages such as isiNdebele.

Prof Langa Khumalo

The second respondent to the keynote address was Prof Leketi Makalela, a founding Director of the Hub for Multilingual Education and Literacies (HuMEL) at the University of the Witwatersrand (Wits). Makalela opened his presentation with a metaphorical presentation of what he called "two elephants sitting in our classrooms". The first elephant is that institutions of higher learning do not prepare new teachers to teach the curriculum utilising new pedagogical teaching practices, and the second elephant was referenced as having ambitious proposals and sound policies that lack proper application. He further asserted that learners and students are disproportionately disadvantaged in two ways: epistemic access (to know and how to know) and identity affirmation (to be). As a result of this, he postulated that monolingualism and epistemic biases are still being normalised.

Dr Sree Ganesh Thottempudi from the Centre for Accounting Studies, part of the College of Accounting Sciences at Unisa, presented as a third respondent under the topic entitled AI for Multilingualism: Demonstrations of AI Tools for Language Teaching and Learning. His presentation focused on developing innovative tools and methodologies to support endangered and low-resourced languages. He presented the creation of linguistic resources, computational models and digital infrastructure tailored to the unique challenges of the Telugu indigenous languages he researched, designing NLP pipelines, language documentation tools, and annotation platforms that empower linguists, communities and researchers.

In reviewing the symposium, Dr Feziwe Shoba from DLS presented the key insights and recommendations in the form of resolutions to chart the future for the directorate and Unisa at large.

Key insights

• Language is central to knowledge acquisition and sense-making.
• Multilingualism is a vast resource and asset for transforming teaching and learning.
• The use of AI is inevitable in the 4IR revolution.
• Emphasis on AI's impact on language learning is that it provides instant feedback and is becoming increasingly accessible.

Recommendations 

• AI as an enabler of multilingualism.
• Evolve with the languages and the infrastructure.
• Low-resourced languages should be digitised and digitalised.
• Languages should acquire digital functions, and their speakers digital skills.

* By Dr Feziwe Shoba and Babalwa Resha, Directorate Language Services