
Audit AI outputs for centralized decision-making patterns in your workflows
You'll have a documented list of 3-5 decision points in your work where AI recommendations show bias toward standardized/uniform solutions, plus notes on where human judgment should override.
What you'll make
AI Centralization Audit
- Decision Authority Map: who currently decides what in your workflow
- Centralization Flags: 3-5 AI recommendations that shift decision authority
- Override Protocols: when to reject AI suggestions in favor of distributed judgment
- Stakeholder Impact: who loses agency when AI centralizes these decisions
Write each section with specific examples from your actual AI-assisted workflows
The 3 parts
Each part ends with a check you can see.
Map your decision landscape
Identify who currently makes key decisions in one AI-assisted workflow you use regularly.
Done when
You have a visual map showing 5-7 decision points in one workflow, noting who decides and what information they use.
Test AI for centralization bias
Use AI on one mapped decision point and document how it shifts authority toward standardized solutions.
Done when
You have tested AI assistance on one specific decision from your map, with notes on how AI's suggestions changed who would decide and what information they'd use.
Design override protocols
Create specific rules for when to reject AI centralization and preserve distributed human decision-making.
Done when
You have 3-4 concrete override rules that specify when to reject AI recommendations that centralize decision authority inappropriately.
Capstone: you put together your AI Centralization Audit.
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
- Audit 3-5 AI-assisted decision points in your work for centralization bias
- Document specific moments where AI pushes standardized solutions over contextual judgment
- Create override rules for when human distributed decision-making should take precedence
Who it's for
- Anyone using AI for workflow optimization, project management, or team coordination
- People who've noticed AI recommendations feeling 'off' but can't articulate why
- Managers wondering when to trust AI suggestions about process standardization
What it won't cover
- Broad AI safety theory or existential risks
- Technical implementation of AI systems
- Workplace organizing or labor negotiations
- Comprehensive bias detection across all AI domains
Skills you'll practise
- Decision authority mapping
- Centralization bias detection
- Override rule development
- Stakeholder impact assessment