Course Papers
Contents
One card per paper, each linking to the canonical source (arXiv, journal DOI, or publisher). Where a paper is paywalled, the DOI is given so you can reach it through the UCT library.
We link to papers rather than hosting copies, so each one stays at its authoritative version. The course materials themselves (lesson HTML and the rest of this site) are CC BY 4.0; that licence does not extend to the linked third-party papers.
Video-led introductory week
Architecture, training, and alignment of large language models
Energy, water, hardware lifecycle, and the rebound problem
Ethics, ubuntu, transparency, and the just-AI framework
Tools, the hallucinated-citation crisis, and AI-augmented research workflows
How AI changes the way we write, think, and generate research ideas
Code generation, data analysis, visualization, verification, and agentic workflows
Vision, documents, audio, and video: what AI can really do across modalities
Benchmarks, failure categories, and where AI is now genuinely strong
Harnesses, long-horizon reliability, MCP, agentic RAG, and the 2026 tool landscape
Drafted with Claude (Anthropic) and reviewed by Jonathan Shock before publication. AI-generated errors are possible — if you spot one, please email jonathan.shock@uct.ac.za. Full detail: AI Content Disclaimer.
© 2026 Jonathan Shock · MAM5020F: Generative AI for Research · CC BY 4.0