Episode Description:
Tom and Christina sit down with AI researcher, Oxford academic and health equity advocate Maxine Mackintosh for a conversation about bias, power and what happens when healthcare innovation grows up.
They explore why the era of "sparkly innovation" in the UK has given way to hard questions about evidence and cost-effectiveness, how AI scribes became one of the first technologies clinicians actively pulled into practice, and why bias in AI isn't just about models — it's about the entire pipeline from data collection to post-market surveillance. From genomics datasets skewed toward European ancestry to language, culture and translation, Maxine makes the case that AI can both amplify inequity and create new opportunities for debiasing. The episode closes on a wider lens: billionaire influence, concentrated power, and why introducing friction — not removing it — might be key to preserving humanity in an automated future.
Key Themes & Topics:
- The shift from "innovation joy" to evidence and economic reality
- Pull vs push adoption: why clinicians are demanding AI scribes
- Bias across the full AI lifecycle
- Genomics data imbalance (85% European ancestry datasets)