Post by University of Washington Information School

🖲️ Ph.D. student Imani Finkley, Informatics grad Yuanxi Li and Assistant Professor Melanie Walsh’s new paper "Neutrality Bites: Gender Representation in AI-Generated Animal Stories" presented their paper at #FAccT2026, the ACM Conference on Fairness, Accountability, & Transparency. 🖥️ The research team examined how large language models (#LLMs) handle gender assignment in stories about talking animal characters. They prompted six leading LLMs to complete an English-language story about seven different animal characters. Across the 23.8K stories, they found that models frequently avoid gendering the animal character in the story (19% on average) or use gender-neutral language like “it” or “its” (38.2% on average). Feminine animal characters are virtually absent, present in just 2.2% of stories vs. 40.6% that feature masculine characters. These findings point to a broader argument: neutrality bites. In other words, models that prioritize neutrality to address social bias may actually contribute to the erasure of marginalized perspectives and identities. Finkley, Li and Walsh suggest that alternative strategies beyond neutrality need to be pursued, such as those that more equally distribute social possibilities across imagined subjects. 💠 Imani Finkley shared this reflection: “This was my first academic conference, first publication, first first-author paper, and first year of my Ph.D. Needless to say, I was excited to be able to connect with other students; learn about the growing literature in the human-AI collaborations and AI policy; and present the first of my doctoral work to my peers. I am happy that FAccT facilitated a dynamic and welcoming community for both myself and our work; I look forward to returning in the future.” 🎉 Congrats to the team! Read the paper: https://lnkd.in/e_xKJc-q