In a rapidly evolving professional environment, the ability to interact effectively with artificial intelligence has become an essential skill. This one‑day training is designed for participants who want to understand generative AI, use ChatGPT (and similar tools) productively, and obtain concrete results as early as the next day. Through a practical and accessible approach, participants will learn to formulate clear requests, structure reusable prompts, refine outputs through targeted iterations, and produce high‑value deliverables (summaries, tables, communications, action plans).
The training also covers key topics such as security, cybersecurity, compliance (including Quebec’s Law 25), and AI governance for responsible use. |
Contents | Introduction and Context Setting - Why AI now: computing power, data, algorithmic advances
- Principles of responsible use: confidentiality, validation, caution
- Selection of practice cases: RFP, municipal document, provided scenario
Understanding AI at a High Level - Everyday examples of AI (recommendations, anti‑spam, fraud detection, translation)
- Supervised vs. unsupervised learning; the “black box” concept
- Generative AI: what it does well, what it struggles with, and why it can make mistakes
ChatGPT in Practice: Interface and Useful Settings - Guided tour: start, find, organize, share, history
- Working with content: excerpts, formats (tables, sections), limitations
- Custom instructions: short and secure version
Collaboration in ChatGPT - Managing teams in ChatGPT
- Collaboration modes (Conversations, Workspaces, Projects)
- Custom GPTs and sharing
Modern, Accessible Prompt Engineering - Core method: frame, break down, structure, write, verify
- Few‑shot prompting: providing 1–2 examples of expected tone or format
- Self‑review: spotting gaps, inconsistencies, risks, validation points
- Iterations: improving clarity, precision, and format through targeted refinement
Case Workshops: Guided Exercise and Breakout Sessions - Guided exercise using a document: extract into a table and generate questions
- Breakout rooms: communication, synthesis, operational tasks
- Submission via chat: 1 deliverable, 1 prompt template, 1 improved iteration
- Data analysis: short guided exercise with a small CSV (structuring, validating, avoiding fabrication)
Security, Cybersecurity, Law 25, and AI Governance - Risks and responsibilities: confidentiality, traces, minimization, anonymization
- Data sensitivity grid: public, internal, sensitive — and associated actions
- Law 25: concrete reflexes for AI use (avoid personal data, anonymize, document practices)
- AI governance: simple rules, permitted/forbidden use cases, validation, escalation
- Practical link: data sensitivity and required level of oversight (personal vs. organizational context)
Final Challenge and Adoption - Final challenge based on a realistic case (RFP, municipal context, data, communication)
- Conclusion: 7‑day adoption plan, mini prompt library, simple performance indicator
|
