MAM5020F — Generative AI for Research
Course Outline
Generative AI for Research: Applications and Implications
Contents
Weekly Structure
Pre-Class Reading materials and videos to prepare for discussion
In-Class Interactive discussions, activities, and collaborative learning
Post-Class Deeper reading and practical assignments to apply concepts
What You'll Be Able to Do
By the end of the twelve weeks, you should be able to:
- ▸Read AI claims calibratedly: take any claim about what AI can do for research and place it honestly as real, overclaimed, or aspirational, and check it against primary sources rather than press releases.
- ▸Use AI across the research lifecycle (literature search, writing and ideation, data analysis and code, and multimodal material) while knowing where each use is genuinely strong and where it quietly fails.
- ▸Verify AI output systematically: catch hallucinated citations, silent errors in an analysis, and confident-but-wrong reasoning, and build the verification into your normal workflow.
- ▸Work ethically and transparently: apply an explicit ethical framework to your own AI use and disclose that use to the standard journals and funders now expect.
- ▸Reason about cost and access: account for the material and environmental footprint of AI, and for the global inequities in who can actually use it.
- ▸Make and defend a judgement: decide where AI does and does not belong in your research, and articulate that position clearly enough to defend it (the Week 12 capstone).
No single week delivers all of these; they accumulate. Calibrated reading runs through the other five and ties them together.
12-Week Journey
Week 1
Foundations of Generative AI
- Historical development and recent breakthroughs
- Distinctions between different AI paradigms
- Introduction to generative AI architectures (transformers, diffusion models)
Week 2
LLM Deep Dive
- How large language models work
- Training, fine-tuning, RLHF and alignment
- How AI image generation works
Week 3
Environmental Implications of AI
- Energy consumption of large model training and deployment
- Carbon footprint calculations
- Sustainable AI practices and green computing
Week 4
Ethical Frameworks for AI in Research
- Transparency and attribution
- Privacy considerations
- Bias, fairness, and representation
- Academic integrity when using AI tools
Week 5
AI-Assisted Literature Reviews
- Free and paid AI literature tools
- The hallucinated-citation problem
- Building a research workflow with Claude
Week 6
AI for Writing, Communication and Research Ideation
- Writing as thinking
- Using AI to generate and test research ideas
- Scientific integrity and an AI writing workflow
Week 7
AI for Data, Code and Computation
- Natural language to code
- AI-assisted data analysis and visualisation
- Verifying AI-generated code
- Agentic data analysis
Week 8
Multimodal AI for Research
- Scientific images and document intelligence
- Transcription and audio analysis
- Video and multimodal workflows
Week 9
Critical Evaluation and Limitations of AI
- The trajectory of LLM capabilities
- Where AI is now strong, and three categories of failure
- Illusions of understanding and verification protocols
Week 10
Agentic AI, RAG and Advanced Research Tools
- What agents are, and how long tasks fail
- Available tools, including MCP
- RAG and a curated tour of research tools
Week 11
Future of AI in Research and Africa's Sovereign AI Capacity
- Possible futures for AI in research
- Policy, peer review and integrity
- Sovereign AI capacity: compute, data, languages, institutions and talent
Week 12
Integrative Capstone
- A 3-hour solo activity completed in your own time
- A 600-word structured research pitch
- A 300-word self-critique of the pitch
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