
Build a verification workflow for AI-generated claims before sharing externally
You'll have a documented process specific to your work that identifies which AI outputs require human verification, secondary sources, or fact-checking before use in communications with others.
What you'll make
AI Output Verification Protocol
- Risk Assessment Criteria: flags that trigger verification
- Verification Steps: what to check and how
- Decision Framework: when to share, edit, or reject AI output
The 3 parts
Each part ends with a check you can see.
Identify Your Surrender Patterns
Audit recent AI-assisted communications to discover where you accepted outputs without verification
Done when
You have a ranked inventory of 5+ recent AI-assisted communications categorized by surrender risk level with specific examples of what you didn't verify
Build Verification Checkpoints
Create specific verification steps for different types of AI outputs based on risk and consequence level
Done when
You have a 3-tier verification system with specific steps for each tier and know which types of AI outputs trigger which level of checking
Test Your Workflow Against Real Stakes
Apply your verification system to upcoming communications and refine based on what actually works
Done when
You have tested your verification workflow on 3+ real communications, documented what worked and what didn't, and have a refined process you'll actually use consistently
Capstone: you put together your AI Output Verification Protocol.
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
- Build a documented verification workflow that flags AI outputs requiring fact-checking before external use
- Distinguish high-risk from low-risk AI outputs using specific criteria you develop
- Test and refine verification steps on your actual work communications
Who it's for
- Professionals who share AI-assisted content with clients, stakeholders, or the public
- Knowledge workers who use AI for research, analysis, or communication drafts
- Anyone who needs to maintain credibility while benefiting from AI assistance
What it won't cover
- How to fact-check specific domains or technical claims
- Building AI systems with uncertainty quantification
- Comprehensive prompt engineering techniques
Skills you'll practise
- Risk assessment of AI outputs
- Verification workflow design
- Evidence quality evaluation
- Cross-source validation methods