The Absent Dataset examines biases embedded in face recognition technologies. Revealing an explicitly unfair dataset of people erased by AI systems, it is based on Twitter’s racist cropping algorithm that prioritizes lighter, slimmer  and younger faces. By highlighting negative aspects of biased algorithms - bluntly exposing the problem of exclusion  and misidentification in face recognition, it illustrates marginalized groups in AI conversations. In a flip of narrative,  it explores aesthetics of absence & gives visibility to groups underrepresented in typical training datasets. The face  tracker’s labels feature positive adjectives, opposing to derogatory terms minorities are labelled as. What’s the meaning of erasure and what are the consequences of incomplete demographics representation.