AI systems are increasingly being used in education, but what happens when those systems introduce hidden bias into learning, feedback, assessment and career advice?
In this episode, we're joined by Victoria Hedlund (AKA The AI Bias Girl), founder of GenEd Labs, former physics teacher, consultant and researcher specialising in AI bias in education.
Victoria shares practical examples of how AI tools can produce different outputs based on gender, names, backgrounds and other personal characteristics. From explanations of how a light bulb works to career guidance and assessment, we explore how seemingly small biases can compound over time and influence attainment, wellbeing and opportunity.
The conversation covers AI tutors, safeguarding concerns, assessment risks, the importance of keeping humans in the loop, and why educators need to adopt a more experimental mindset when working with generative AI.
This is a thought-provoking discussion about fairness, agency and what every teacher needs to know as AI becomes increasingly embedded in education.
For more background, here's two bits of Victoria's writing:
Deep dive on using AI for marking
https://victoriahedlund.substack.com/p/bias-bytes-special-edition-evidence
Deep dive on Squishmallow Top Trumps
https://www.linkedin.com/posts/victoriamhedlund_biasgirl-biasaware-chatgpt-ugcPost-7452959849623220225-bIWG
And finally, free downloads from Victoria, including 8 bias-mitigating prompts for teachers, and a growing set of other resources on AI bias in education.
https://genedlabs.ai/downloads
Ireland