LLM Hallucination

The phenomenon where large language models generate content that is factually incorrect, fabricated, or not grounded in the input provided, while presenting it with apparent confidence. In bias research, hallucinations are particularly concerning because models may invent fictional narratives that reinforce stereotypes—for example, generating elaborate stories about older adults struggling with technology when the prompt contained no such information. LLM hallucinations can amplify social biases by creating detailed, convincing-sounding scenarios that embed stereotypical assumptions about age, disability, race, or gender. Understanding hallucination as a vector for bias propagation is important for accessibility practitioners who design AI-powered tools for marginalized communities.