Algorithmic Bias

Systematic and unfair discrimination embedded in the outputs of algorithmic systems, arising from biased training data, flawed model design, or unrepresentative development processes. For people with disabilities, algorithmic bias manifests in multiple ways: voice assistants failing to recognize atypical speech, hiring algorithms screening out disabled applicants based on facial analysis or personality assessments, content moderation systems suppressing disabled creators, and predictive systems misinterpreting disability-related behaviors as suspicious or fraudulent. Unlike demographic biases based on race or gender, disability bias is often invisible because many disabilities are not captured in training data.