
Create a decision checklist that forces verification before accepting AI outputs
You'll have a 4-6 point verification protocol you apply to every significant AI recommendation before acting on it in your work.
What you'll make
AI Output Verification Checklist
- Context Check: does AI understand what I'm actually trying to solve
- Edge Case Review: where might this recommendation break down
- Stakes Assessment: what happens if this is wrong
- Override Rules: when to ignore AI and trust your judgment
- Quick Tests: how to validate AI output in under 2 minutes
Write each checkpoint as a yes/no question you can answer quickly before acting on AI advice
The 3 parts
Each part ends with a check you can see.
Test AI on One Real Decision
Pick one actual decision from your work this week and see where AI gets it right vs wrong.
Done when
You have AI's recommendation for a real decision plus your own assessment of where it's naive or missing context.
Map Your Personal AI Failure Patterns
Test AI on 3 different types of your work to discover where it consistently misleads you.
Done when
You have 3 examples of different AI failure modes and can name the pattern in each one.
Build Your Verification Protocol
Create a 4-6 point checklist that catches AI failures before they become your mistakes.
Done when
You have a working checklist with yes/no questions that take under 2 minutes to answer.
Capstone: you put together your AI Output Verification Checklist.
Where this helps
An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.
About this lesson
What you'll be able to do
- Apply a 4-6 point verification checklist before acting on any significant AI recommendation
- Distinguish between AI outputs you can use immediately vs those requiring human oversight
- Identify the 3 most common failure modes where AI misleads you in your specific work context
Who it's for
- Anyone using AI for work decisions who's wondered when to trust vs verify AI output
- People who've been burned by acting on AI recommendations that seemed solid but weren't
- Beginners who want a systematic approach to AI use rather than trusting their gut
What it won't cover
- Technical AI model evaluation or bias detection methods
- Industry-specific compliance requirements for AI use
- Advanced prompt engineering techniques
Skills you'll practise
- Systematic verification of AI recommendations
- Risk assessment for AI-assisted decisions
- Personal standard articulation
- Failure pattern recognition