Bias We Have Seen — Disability Representation in Digital Tools
One example of bias in digital technology is the limited representation of persons with disabilities in AI-generated images and online educational content. For example, when asked to generate an image of a classroom in Nepal, an AI image generator may produce a classroom where all students are non-disabled, even though students with physical, visual or other disabilities are also part of our schools.
Affected dimension: Disability and inclusion.
Why is it unfair: Such representation reinforces the assumption that disability is uncommon or that persons with disabilities do not belong in ordinary classrooms. It can make students with disabilities feel invisible and strengthen existing social stereotypes. This bias may arise because the training data does not adequately represent diverse learners and learning environments.
Practical response as a teacher: Teachers can deliberately use diverse and inclusive learning materials, including images and examples that represent persons with different disabilities as active learners and participants. They can also question AI-generated content, review it for stereotypes and exclusion, and encourage students to critically examine whose perspectives are missing.