Abstract
This study examines various factors that influence the adoption of Artificial Intelligence (AI) in STEM teaching for first-year undergraduate students at a private higher education institution in South Africa. Grounded in the Technology Acceptance Model (TAM), the research utilises a qualitative approach, including interviews with lecturers, focus groups discussions, and classroom observations to capture the lecturers' perceptions, challenges, and strategies in AI integration. Key findings reveal diverse integration patterns, from sporadic to structured use of AI tools such as ChatGPT, with notable benefits in enhancing student engagement, facilitating complex problem-solving, and streamlining lesson planning. However, challenges such as student misuse, technical limitations, and insufficient institutional support were evident. Lecturers demonstrated adaptability by leveraging AI-generated content as teachable moments, guiding responsible use, and relying on peer networks for informal support. The study underscores the importance of tailored institutional policies, continuous training, and community-driven learning networks to foster effective AI adoption in STEM teaching. Insights gained contribute to the wider discussion on AI's role in transforming educational practices and highlight pathways for improving AI integration through targeted support and adaptive strategies.
Keywords: Artificial Intelligence (AI), STEM Education, Technology Acceptance Model (TAM), Higher Education, AI Integration, Lecturer Adaptation