AI Literacy and Responsible Generative AI at Work
Write better prompts, verify outputs, protect confidential data and watch for bias when using generative AI.
Content changes frequently. AI tools and workplace policies change rapidly. Follow your organisation's current AI policy.
Why it matters
AI can speed up report drafting, checklists and training material — but it can also invent facts and leak confidential information if used carelessly.
Who it's for: All professionals using AI chat tools for work.
Learning objectives — by the end you can:
- Write a clear prompt with role, task, context and format.
- Verify AI output against reliable sources.
- Decide what must never be entered into public AI tools.
Key terms
- Hallucination
- Confident but false output from an AI model.
- Prompt
- The instruction you give the AI.
- Human in the loop
- A person reviews and is accountable for the final output.
Lesson notes
Prompt pattern
| Part | Example |
|---|---|
| Role | You are a safety trainer |
| Task | Draft a 5-minute toolbox talk |
| Context | Topic: ladder safety for warehouse staff |
| Format | Bullet points, plain English, 200 words |
Never enter into public tools
- 1Personal data of colleagues, visitors or customers.
- 2Security layouts, camera positions, access codes or incident details.
- 3Confidential contracts or unreleased financials.
Verify
Check facts, numbers, standards and legal references against official sources. You are accountable for what you publish.
Video lesson
Video to be added
A verified video from an official or reputable channel has not yet been selected for this topic. The text lesson above covers the full content.
Search YouTube for related videosPractical story: The invented standard clause
Fictional teaching example
Situation. A supervisor asked an AI tool to cite the clause of a standard requiring a specific inspection interval.
Known facts
- The AI produced a plausible clause number and wording.
- On checking, the clause did not exist.
Decisions
- The supervisor used the organisation's licensed copy of the standard instead.
Outcome. The procedure cited the correct requirement.
Lessons
- AI can fabricate references.
- Verify against the source.
- Use AI for drafting, not authority.
Worked example
Worked example: improving a prompt
| Weak | Better |
|---|---|
| 'Write about fire safety' | 'As a fire warden trainer, write a 1-page evacuation checklist for a 3-floor office, numbered steps, plain English.' |
Practice exercise
Write a prompt to draft a visitor safety briefing card. Then list two things you would verify.
Quiz
Sign in to save quiz results.
References
- Official guidanceNIST — AI Risk Management Framework
- Official guidanceUK NCSC — AI and cyber security
Verify current editions and applicability. Standards are copyrighted; consult licensed copies for exact requirements.
Report an issue with this contentRelated topics
Cybersecurity Awareness
Cybersecurity Awareness: Phishing, Passwords and MFA
Spot phishing, use strong unique passwords with a manager, turn on MFA and report incidents fast.
Privacy & Data Protection
DPDP Act, 2023 — Awareness for Workplaces
Understand the core ideas of India's Digital Personal Data Protection Act, 2023: notice, consent, purpose limitation, minimisation, rights and breach readiness.