
Build a verification system to catch AI errors before they reach your work output
You'll have a 3-step checklist you use before sharing or acting on any AI-generated content, plus a log of errors your AI tool commonly makes in your domain.
What you'll make
AI Verification System
- Pre-Use Checklist: 3 verification steps for any AI output
- Error Pattern Log: Common mistakes your AI tools make
- Risk Thresholds: When to verify vs when to trust
The 3 parts
Each part ends with a check you can see.
Test AI Against Known Facts
Feed your AI tool information you already know to see where it gets things wrong
Done when
You have documented 3+ specific error types your AI tool makes with examples of each
Design Your 3-Step Checkpoint
Create a quick verification process you'll actually use before sharing or acting on AI output
Done when
You have a 3-step checklist that takes under 5 minutes and catches the errors that matter most in your work
Test Your System on Real Work
Use your 3-step checklist on actual AI outputs from your work and refine what doesn't work
Done when
You have used your verification checklist on 3+ real AI outputs and refined it based on what you discovered
Capstone: you put together your AI Verification System.
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
- Catch AI errors before they impact your work decisions
- Run a 3-step verification process on any AI output you plan to use
- Build and maintain a personal error pattern log for your AI tools
- Distinguish between AI outputs you can trust vs those requiring extra validation
Who it's for
- Knowledge workers who use AI for analysis, writing, or decision support
- Professionals who've been burned by AI mistakes and need better quality control
- Anyone who shares AI-assisted work with colleagues or clients
What it won't cover
- How to prevent AI from making errors in the first place
- Technical prompt engineering techniques
- AI model comparison or selection
- Domain-specific fact-checking methods
Skills you'll practise
- Output verification workflow design
- Error pattern recognition
- Quality control checkpoint implementation
- Risk-based validation prioritization