Abstract:

Introduction:
In South Africa, multilingualism is an important aspect of deepening democracy and inclusion in both higher education institutions and in post-apartheid society. Since the dominance of English in higher education both reflects and perpetuates colonization, higher education institutions look to language policies to promote decolonised teaching and learning. According to the Revised Language Policy for Higher Education this entails strengthening language use to include the use of indigenous languages for communication, scholarship, teaching, and learning. Implementing the policy requires teaching staff in higher education institutions to deepen their language awareness and broaden their use of languages in teaching and learning.
Despite students, preference for learning English, they benefit from having access to African languages for teaching and learning. Bilingual language policies have a social value but need to be balanced against the pressure of globalization that privilege imported languages over indigenous ones. This tension can be managed by using multilingual pedagogies. Similarly, intellectualization of indigenous languages is necessary support multilingualism in tertiary institutions.
The envisioned policy has several implications for language pedagogy: they include broadening the range teaching approaches used for indigenous languages, producing appropriate language learning materials and supporting multilingual scholarship. This study explores the potential and limitations of GenAI (Generative Artificial Intelligence) in meeting these challenges. It uses a case study of teaching conversational Northern Sotho to tertiary students.

Methodology:
The methodology uses GenAI to produce lesson plans and learning activities that reflect each of three alternative approaches to traditional language teaching. The generated text is analysed in terms of language accuracy, the degree to which generated lessons reflect alternative learning approaches and suitability for meeting the needs in tertiary students.

Findings:
The findings show that Gen-AI makes a useful contribution to expanding the range of approaches to Northern Sotho language teaching by suggesting appropriate activities that reflected three different language teaching approaches. However oral activities that used pronunciation tended to use English rather than Northern Sotho phonetics. The activities however also rendered some common grammatical errors.
These findings highlight the important role lecturers play in thoughtfully mediating language learning with genAI, in ensuring accurate learning that meets the evolving demands of higher education.

Keywords:

African language teaching, Generative artificial intelligence, Approaches to language teaching.