
Document your assumptions about what AI can safely do in your role
You'll have a written list of tasks where you assume the AI is 'safe enough' and the reasoning behind each assumption—creating an audit trail for decisions and a baseline to revise as threats evolve.
What you'll make
AI Safety Assumption Audit
- Task-Safety Map: where you assume AI is safe enough
- Drive-Risk Assessment: what training drives create risk
- Revision Triggers: when to update assumptions
The 3 parts
Each part ends with a check you can see.
Map Your Safety Assumptions
Document which of your tasks you currently assume AI can handle safely and why.
Done when
You have a ranked list of 5-7 work tasks with explicit safety assumptions and confidence levels for each.
Test Drive-Risk Mismatches
Research what your AI tools were optimized for and identify where those drives conflict with your safety needs.
Done when
You have identified at least 2 drive-risk mismatches where AI optimization goals create safety concerns for your tasks.
Set Revision Triggers
Define specific conditions that should prompt you to revisit your safety assumptions.
Done when
You have 3+ specific, observable triggers that would prompt safety assumption revision, with a plan for checking each.
Capstone: you put together your AI Safety Assumption 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
- Create a written safety baseline for AI use in your role that you can defend to stakeholders
- Distinguish between tasks where AI uncertainty matters vs. tasks where it doesn't
- Build a revision system for updating safety assumptions as AI capabilities change
Who it's for
- Professionals in roles with safety, compliance, or quality requirements
- Team leads who need to set AI use policies for their group
- Knowledge workers in regulated industries or client-facing roles
What it won't cover
- Technical AI alignment theory
- How to train AI systems
- Organization-wide AI governance policies
Skills you'll practise
- Risk assessment with incomplete information
- Safety assumption documentation
- Cross-scenario stress testing
- Evidence-based boundary setting