Building a shared approach to checking bias in AI lesson plans
Chair’s engagements · Stories & activities
We reviewed prompts, self-checklists and human review criteria to help reduce bias in teaching materials.
Ask for a better image.
Move from spotting bias to rewriting the prompt.
- State the purpose.
- Challenge stereotypes.
- Vary the roles.
- Focus on contribution.
- Activity year
- 2026
- Organiser
- Seoul Foundation of Women & Family
- Role
- Chair Kim Ye Eun · Expert adviser for the Safe Filter research project
- Supporting materials
- Chair’s activity record · Advisory project materials reviewed
Advising on the Safe Filter research project
KAAA chair Kim Ye Eun contributed expert advice to the Seoul Foundation of Women & Family’s research on AI teaching materials and a Safe Filter from a gender-equality perspective.
Questions we explored together
- How can gender stereotypes and role bias be reduced in AI-generated images and copy?
- How should standard prompts and self-checklists be structured for use in childcare and education?
- What can automated AI checks do, and what still needs human review?
The review focused on prompt design, assessment criteria, practical use and the limits of automated AI checking. It also covered training and workshop delivery for the monitoring group.
The hands-on AI gender-equality monitoring training is documented as a separate educational activity.