In this episode of the UKAI Women in AI Podcast, host Zahra Shah sits down with Sarah Wyer, a leading PhD researcher specialising in gender bias in large language models. Sarah shares her non-linear journey into AI, from starting out in business and higher education, to discovering a passion for data, completing a master’s in computer science, and ultimately becoming an award-winning woman in tech. She explains how her research with Professor Sue Black revealed deeply embedded gendered and intersectional bias in early GPT models, and why the data choices made by AI developers profoundly shape societal outcomes. Her guiding principle is simple but powerful: technology must not leave people behind.

Sarah also reflects on the challenges she has faced: from access barriers in AI research to the urgent safety concerns arising from generative models, and outlines her current work mapping gender bias across GPT iterations up to GPT-5. She speaks candidly about the risks posed by sexualised AI systems, the importance of robust governance, and her belief that AI can still be a force for equality if society chooses to embed the right values. Looking to the future, she urges young girls and women not to be deterred from entering the field, highlighting the need for visible role models and supportive communities. The episode closes with Sarah’s call for collective action: to design AI that elevates people, protects vulnerable groups, and creates a more equitable digital world for everyone.

AI is our Business.

UKAI is the Trade Association for AI businesses across the UK. Join us, ukai.co