Lesson status
Not completed
Mark lesson complete
Not completed
Saved only in your browser. Return to the course index to see overall progress.
Lesson 3 • Leadership
Bias, Equity, and Harm
AI systems can reinforce historical inequities if used without scrutiny. Community organizations are often closest to the people most affected.
Key concepts
- Data can encode past discrimination and unequal access.
- “Efficiency” can erase nuance and lived experience.
- Safeguards must be explicit before adoption.
Practice Exercise
Identify one population you serve, one way biased data could disadvantage them, and one safeguard you would require before using AI.
Template (copy/paste)
ROLE: You are my AI assistant. GOAL: Help me assess bias risks for an AI use case. INPUTS: Population served, data sources, decision impact. OUTPUT: 5 risks + 5 mitigations + a 'do not deploy' threshold. CONSTRAINTS: Prioritize equity and explain tradeoffs plainly.
Ethics & accuracy: verify important facts, avoid sharing sensitive personal data, and be transparent when AI helped draft content.