Working with AI Tools

This module teaches learners to use large language models and other AI tools responsibly and effectively in day-to-day work. It covers core concepts such as prompts, context windows, temperature, hallucination, system prompts, few-shot prompting, chain-of-thought, retrieval-augmented generation and fine-tuning. Learners also study prompt injection, data-privacy risks, verification habits and the economics of tokens. Case studies focus on writing, coding, data analysis and research, with ethics woven throughout.

Learning Outcomes

Explain what a large language model does and why it can hallucinate.

Write clear, structured prompts using role, task, context and format sections.

Apply few-shot and chain-of-thought techniques to improve answer quality.

Judge when retrieval-augmented generation is more appropriate than fine-tuning.

Identify prompt injection and data-leakage risks in AI workflows.

Verify AI outputs against authoritative sources before relying on them.

Estimate token costs and context-window limits for common tasks.

Apply professional ethical standards when using AI at work.

Topics Covered

What an LLM is: tokens, probabilities and training data

Prompts, system prompts and user prompts

Context windows and memory limits

Temperature, top-p and other sampling parameters

Hallucination: causes and mitigations

Few-shot prompting and in-context learning

Chain-of-thought and step-by-step reasoning

Retrieval-augmented generation (RAG)

Fine-tuning vs prompting vs RAG

Prompt injection and jailbreaks

Data privacy and confidentiality with AI vendors

Verification, citation and source checking

Token economics and cost estimation

Multimodal AI (vision, audio, code)

Ethics, bias and responsible AI use

Skills Developed

Writing effective prompts for business tasks Choosing between prompting, RAG and fine-tuning Spotting and recovering from hallucinations Protecting private data when using AI Estimating and controlling AI usage cost Combining AI output with human review Explaining AI decisions to non-technical colleagues Applying ethical judgement to AI use cases

Suggested Resources

{'title': "Anthropic's Prompt Engineering Guide", 'author': 'Anthropic', 'source': 'docs.anthropic.com'}

{'title': 'OpenAI Platform Documentation', 'author': 'OpenAI', 'source': 'platform.openai.com/docs'}

{'title': 'Prompt Engineering Guide', 'author': 'DAIR.AI', 'source': 'promptingguide.ai'}

{'title': 'Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks', 'author': 'Lewis et al. (Meta AI)', 'source': 'arxiv.org/abs/2005.11401'}

{'title': 'AI Snake Oil', 'author': 'Arvind Narayanan and Sayash Kapoor', 'source': 'Princeton University Press'}

{'title': 'The Alignment Problem', 'author': 'Brian Christian', 'source': 'W. W. Norton'}

{'title': 'UK AI Safety Institute Research', 'author': 'AISI', 'source': 'aisi.gov.uk'}

Assessment

Learners complete a prompt library for a chosen workplace scenario, write a short RAG design note, perform a hallucination audit on a provided transcript and sit a final MCQ test.

prompt_library20%

rag_design_note20%

hallucination_audit20%

final_mcq_test20%

ethics_reflection20%

Practice bank: 200 MCQs · Timed mock & self-paced modes

Start Practice →