A new University of Washington study suggests that artificial intelligence models exhibit gender bias in children's animal stories. When six leading AI models were asked to complete short stories about animals, female characters appeared only 2% of the time. Male characters accounted for 41%, while the remaining 57% were presented as gender-neutral or without any explicit gender. The findings were presented June 25 at the 2026 ACM Conference on Fairness, Accountability, and Transparency in Montréal. The research builds on earlier work by Melanie Walsh, an assistant professor in the University of Washington’s Information School, who examined 300 popular children’s books with The Pudding, finding that most frequently appearing animals were described as male, with cats, ducks and birds slightly more often female, and frogs and wolves having more than a 90% chance of being referred to as “he.” In the new study, researchers submitted variations of the same unfinished-story prompt to Claude Sonnet 4.5, Gemini 2.5, GPT-4o, GPT-5.1, Mistral Medium and Olmo 3, producing 23,800 responses. They tested seven animals—bears, birds, cats, dogs, mice, pigs and rabbits—in four settings. Neither temperature nor setting substantially changed the overall gender distribution. Cats were assigned female identities 7% of the time, the highest proportion recorded. Birds were neutral in 96% of responses. The models frequently avoided gendered pronouns, using “