
Establish a verification protocol before accepting any AI-generated decision or design
You'll have a pre-implementation checklist specific to your domain that you run every AI recommendation through, plus documented instances where scrutiny caught flawed outputs.
What you'll make
AI Verification Protocol
- Pre-Implementation Questions: Your domain-specific red flags
- Verification Actions: What you actually check
- Failure Case Library: Real examples where scrutiny caught problems
- Cognitive Load Signals: When you're surrendering vs offloading strategically
Submit your checklist with at least 3 real examples of AI outputs you initially accepted but later caught as flawed
The 3 parts
Each part ends with a check you can see.
Map Your Domain's AI Failure Modes
Identify the specific ways AI goes wrong in your type of work by testing it on real tasks
Done when
You have 5+ documented examples where AI gave plausible but wrong answers in your domain, with patterns identified
Design Domain-Specific Verification Gates
Build a pre-implementation checklist that catches your documented AI failure patterns
Done when
You have a 4-7 item checklist that would catch the failure modes you documented, tested on at least 2 new AI outputs
Monitor Your Cognitive Surrender Signals
Track when you're scrutinizing AI less carefully and build triggers to re-engage verification
Done when
You can identify 3+ personal signals that you're accepting AI output without adequate checking, with tested interventions for each
Capstone: you put together your AI 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 domain-specific verification checklists that catch AI reasoning flaws
- Document and learn from AI output failures in your actual work
- Distinguish between helpful AI assistance and cognitive surrender
- Maintain deliberate oversight when AI handles complex decisions
Who it's for
- Knowledge workers who use AI for analysis, recommendations, or design decisions
- Professionals whose work quality depends on catching errors before implementation
- People noticing they scrutinize AI output less carefully over time
What it won't cover
- How to build AI systems
- General prompt engineering techniques
- Theoretical frameworks about attention and cognition
- AI safety at organizational scale
Skills you'll practise
- Risk-based verification design
- Critical output analysis
- Cognitive load management
- Evidence-based decision validation