Uncovering Bias in AI

In this episode of Trailblazing Schools, Trudy explores the biases inherent in AI and large language models, drawing parallels with historical biases in publishing and education. She emphasises the importance of understanding these biases to use AI responsibly and ethically.Trudy discusses the impact of bias on marginalised communities, strategies for recognising and mitigating bias and the importance of diverse training data to reduce bias in AI. Finally, she highlights the role of education in addressing bias and leaves the listener with some practical questions to reflect on.Resources:Find out more about the Trailblazing Schools Partnership here: https://www.trudygraham.com.au/trailblazing-schools/ Find out more about the Leaders Learning Circle here https://www.trudygraham.com.au/leaders-learning-circle/ Get your copy of AI in Action: Field Notes from School Leaders. It's free, practical, and designed to prompt your own next step. Download today at trailblazingschools.com/fieldnotes If this conversation resonated and you're exploring how to strengthen leadership in your school network or system, visit trudygraham.com.au/connect  to explore working together. CreditsThe Trailblazing Schools Podcast is hosted by Trudy Graham - https://www.trudygraham.com.au/This podcast is edited and produced by Ellen Ronalds Keene from PERK Digital - https://perkdigital.com.au/The theme music is Dusty Road by Indiebox Trudy Graham acknowledges the traditional custodians of the land on which she lives and is recording today, the Darumbal people, and pay my respect to elders past and present, for they hold a deep and continuing connection to the land, sea, and sky.Chapters00:00 Introduction: Bias in AI and Education01:56 Historical Bias in Data and Publishing03:55 Representation Bias and Its Consequences05:00 Publication Bias and Its Impact on Knowledge06:07 Naming AI to Foster Critical Thinking07:07 Bias Beyond Representation: What's Not Seen08:04 Historical Power and Marginalized Voices09:02 The Matilda Effect and Gender Bias10:00 Bias in Publishing and Its Legacy10:58 Practical Steps to Recognize Bias in AI12:09 Encouraging Curiosity and Critical Inquiry12:57 Closing Thoughts: The Role of Education and InquiryAI bias, large language models, representation, historical bias, education, ethical AI, diversity in publishing, bias in training data