Artificial Intelligence Research: 6th Southern African Conference, SACAIR 2025, Cape Town, South Africa, December 1–5, 2025, Proceedings(2784 Communications in Computer and Information Science)

This book constitutes the refereed proceedings of the 6th Southern African Conference on Artificial Intelligence Research, SACAIR 2025, held in Cape Town, South Africa, during December 1–5, 2025. The 36 full papers presented in this volume were carefully reviewed and selected from 150 submissions.The papers are organized into the following topical sections: Algorithmic and Data Driven AI; Socio-technical and Human-Centred AI; and Responsible and Ethical AI.  

Table of Contents:
.- Algorithmic and Data Driven AI. .- Which contextual topic modelling algorithm is the best?. .- Leveraging Language Models for Document Type Classification in  Low-Resource Afrikaans Archives. .-  Hybrid Automatic Modulation Classification for Increased Robustness  under White-Box Adversarial Attacks. .- Counterfactual Explanation Model for Personalised Dietary  Interventions in Anaemia Patients. .- Feature extraction for plant growth estimation. .- Exploring the impact of different loss functions for anomaly prediction  in a mineral processing plant. .- Modifying class distributions to improve the classification of minority  group examples in a class-imbalanced dataset. .- Evaluating OCR performance on food packaging labels in South Africa. .- Investigating the Impact of Multilingual Pre-trained Speech Models on  Gender Bias in ASR for Low Resource African Languages. .- A Two Stage Pipeline for Automated Caries Detection on Single Tooth  Images from Panoramic Radiographs. .- Exploring syllable similarity across South African languages through  self-supervised speech representation. .- Data Pruning: Redundant, Problematic, and Interdependent Samples. .- Restoring Neural Network Plasticity for Faster Transfer Learning. .- AfroXLMR-Comet: Multilingual Knowledge Distillation with Attention  Matching for Low-Resource Languages. .- Code-Switch Pretraining for Improved Cross-Lingual Alignment in  Low-Resource Languages. .- Identification of Social Media Users that Perpetuate Xenophobic  Attitudes and Hate Speech Narratives in South Africa. .- HAPI-FedProx: Heterogeneity-Aware Adaptive Proximal Optimization  for Federated Learning. .- Fog-Based Deep Learning for Real-Time Cold Chain Temperature  Prediction Using IoT Data. .-  Robust Driver Distraction Recognition via Lightweight Body-Part  Association and Object Context on NVIDIA Jetson. .- Nearest-Class Mean and Logits Agreement for Wildlife Open-Set  Recognition. .- Semi-Supervised Object Segmentation via Active Learning for Efficient  Ecological Monitoring. .- Scaling behavior of Encoder Language Models in Low-Resource Settings. .- Enhancing Target Re-Identification via Model Fusion and  Knowledge Distillation of Pre-trained Foundation Models. .-  Heart Disease Prediction: A Comparative Study of Optimizers’  Performance in Deep Neural Networks. .-  Socio-technical and Human-Centred AI. .- AI-Enabled Cybersecurity Implementation: A Case Study of Critical  Success Factors in a South African SOE. .- Large Language Model Evaluation: From Model-Centric Benchmarks  to System-Level Metrics. .-  Exploring Postgraduate Students’ Use and Perceptions of Generative  AI as a Research Support Tool: A South African Case. .-  An Awareness Framework for Sustainable Selection of LLMs in Business. .- Synthesizing Insights from AI in Africa: A Comprehensive Qualitative  Analysis. .- Factors influencing the use of Autonomous Vehicles on Traditional  Road Infrastructure in The Western Cape. .- An Agentic LLM-Powered Framework for Cholera Risk Prediction and  Actionable Insights from Explainable ML and Statistical Insights. .- A Scoping Review on Generative AI prompting to optimize the  workflow of healthcare professionals in sub-Saharan Africa. .- Detection Of Power Line Insulator Defects Using YOLOv10-N. .-  Responsible and Ethical AI. .- Artificially Generated Minorities (AGMs): The Veneer of Algorithmic  Bias Correction. .- Under-resourced and over-reseached: The ELSA approach to  transdisciplinary AI research in low-trust neighborhoods. .- From Principles to Practice: Extracting Human-Centred AI  Components from Governance Frameworks.