The Artificial Human Why is AI erasing people?
Oct 1, 2025
Guest Jessica Smith, a former Paralympic swimmer and disability advocate, shares her struggles with AI-generated images accurately reflecting her experience of disability. She emphasizes how AI often perpetuates societal biases and the urgent need for inclusive design. Abraham Maldonado, an AI prompt engineer, discusses his work on creating diverse and representative AI outputs. Together, they explore why marginalized communities risk erasure in AI, the importance of diverse prompts, and a hopeful outlook for more inclusive technology.
AI Snips
Chapters
Transcript
Episode notes
AI Failed To Reflect A Paralympian
- Jessica Smith tried to generate AI images that reflected her missing left arm and the models repeatedly showed two arms or unsuitable prostheses.
- She prompted very specifically but the systems still could not produce accurate representations of her body.
AI Mirrors Online Inequality
- AI reflects the biases and inequalities present in its training data because it learns from the internet.
- Missing or deprioritized groups online get erased or misrepresented by generative models.
Representation Shapes Self Image
- Representation affects self-image and future aspirations, especially for children and marginalized groups.
- Lack of visible diversity can internalize stigma and make people feel invisible or less important.