Anonymous note
Your workplace judgement is still the asset.
AI can draft, summarise, compare, and organise. You still decide what is true, appropriate, useful, safe, and worth sending.
Why this matters
The more AI can produce, the more judgement matters.
When output becomes cheap, the valuable skill is not producing more words. It is setting intent, recognising quality, checking risk, and knowing what the work is really for.
The five human checks AI does not own.
1
Intent
What outcome are we actually trying to create?
2
Context
What does AI not know about the people, politics, history, or constraints?
3
Standards
What does good work look like here, for this audience, in this situation?
4
Risk
What could go wrong if this is wrong, misunderstood, or over-trusted?
5
Responsibility
Who owns the final decision after AI has helped?
Copy-paste prompt: keep judgement in the loop.
# ROLE
You are my human judgement checker.
# TASK
Help me review this AI-assisted work before I use it.
# WORK PRODUCT
[paste non-sensitive draft, summary, plan, or analysis]
# CONTEXT
[audience, decision, risk level, what I am responsible for]
# OUTPUT FORMAT
1. What AI helped with
2. What still requires human judgement
3. Missing context or assumptions
4. Quality issues to fix
5. Risks to check
6. Final decision questions I should answer myself
# RULES
- Do not make the final decision for me.
- Be specific about what a human must own.
- Mark uncertainty clearly.
A calmer way to think about AI at work.
AI can help
- Draft options.
- Summarise notes.
- Compare alternatives.
You still own
- Intent.
- Context.
- Final judgement.
Practise
- Ask better questions.
- Check outputs.
- Notice small wins.