Overview: This short lesson and activity supports youth in identifying bias in common AI, such as Google search + image, generative AI, and face detection. The below slides support learning the basic definition of AI, grounded in common AI discussions. Then learners are shown different AI examples that may contain bias, which they first consider individually and then discuss together.
Audience: Middle school-aged youth, can be used with families to support intergenerational learning and discussion
Notes:
These materials are meant to begin conversation about bias in AI and could be paired with more on training data (e.g., see the Data Detectives game) and later design of responsible AI systems.
We found that there's value in having learners both consider individual and then discussing together. Varying views from fellow learners can support more complex understanding of nuanced ethical challenges in AI.
Research Publication:
The Potential of Diverse Youth as Stakeholders in Identifying and Mitigating Algorithmic Bias for a Future of Fairer AI (CSCW '23) ⭐ Recognition for Contribution to Diversity and Inclusion
Augmenting Youths' Critical Consciousness Through Re-design of Algorithmic Systems (ICER '24)
At a high level, learners consider how AI may treat groups of users differently
Learners are shown examples of AI and consider if they have (unfair) bias
This slide deck covers the basics of AI and facilitates an activity, in which learners individually consider and then discuss together if different examples of AI (e.g., search suggestion, generative AI output, face detection) have bias. A supplemental answer sheet may be created to support individual deliberation before group-wide discussions.
Bias in AI.pdf