An edge case refers to a situation, user, or input that falls outside the “typical” or “expected” range that a system is designed to handle. It means groups of people are treated as unusual or rare — and often ignored in design and testing. A speech recognition AI might be trained mostly on nondisabled speakers, treating people with atypical speech patterns as an outlier problem rather than designing inclusively from the start. Hiring algorithms may be optimized for candidates with “standard” work histories, treating disabled applicants with non-linear trajectories as exceptions.