A peer-reviewed study led by Special Olympics with Oregon State University finds that five widely used AI chatbots consistently generate biased stories about people with intellectual disabilities. Researchers had the systems write 25,000 short stories about everyday situations, such as crossing a street, playing sport or going to a store, then compared how people with and without intellectual disabilities were portrayed. Stories about people with intellectual disabilities were far more likely to be paternalistic, depict them as younger and more dependent on others, and cast them as sources of inspiration rather than as people simply going about ordinary lives.
Why it matters: As AI tools increasingly shape hiring, healthcare and education decisions, stereotypes built into their outputs can quietly reinforce real-world discrimination against people with intellectual disabilities.
Source: Special Olympics →
Nextrability 