Episode Summary
Tracey Spicer is one of Australia’s most respected journalists and the author of Man-Made: How the Bias of the Past Is Being Built into the Future. In this episode, Georgie sits down with Tracey for a sharp, funny, and occasionally jaw-dropping conversation about what happens when we treat AI like neutral math instead of what it really is: opinion written in code.
They unpack why algorithmic bias is getting worse in the generative AI era, how recommendation engines can quietly radicalise people (from Andrew Tate pipelines to hyper-performative “tradwife” culture), and why “move fast” without guardrails is a dangerous blueprint. The discussion also goes into the weird and unsettling frontier of humanoid home robots, privacy risks in always-on devices, and what Tracey learned researching sex robots, including the disturbing ways consent is engineered out of the product.
Plus: why Tracey’s favourite AI tool is Claude, what she thinks about Grok and the chaos machine of X, why we are not getting a four day work week anytime soon, and her case for “regulatory sandpits” to test AI safely before it hits the rest of the world.
Chapters:
01:10 – Tracey’s TEDx “The lady stripped bare” moment and why it still matters
04:45 – Beauty standards, AI filters, and why expectations on young women have intensified
08:20 – Man-Made and the epiphany that sparked Tracey’s AI obsession
11:10 – The AI arms race, speed, and why we are in the “seatbelt era” of tech
14:30 – Digital natives vs critical thinking: the hallucination blind spot
16:45 – Tracey’s AI stack: why Claude is her daily driver
19:05 – Humanoid home robots: convenience vs surveillance
21:55 – Strength vs security: what actually scares Tracey about robots
24:35 – Sex robots and the consent problem manufacturers do not talk about
28:10 – Algorithms as “opinions in code” and how radicalisation happens
33:10 – Removing bias: conversations, perspective checks, and inclusive design
35:00 – Grok, MechaHitler, and what happens when platforms mirror their owners
36:45 – Deepfake porn, consent, and why regulation is finally catching up
38:10 – No, AI will not magically deliver a four day work week
41:10 – Future jobs: law, AI assistants, and why juniors still need fundamentals
44:15 – Indigenous knowledge, language revitalisation, and the full-circle AI story
46:50 – Rapid fire: brain chips, Waymo, smart glasses, and AI “snog marry avoid”
49:55 – What we should do now: regulatory sandpits and real guardrails
