GAEA Talks

This week on GAEA Talks, Graeme Scott sits down with Professor Hany Farid, one of the founding figures of digital forensics and one of the most authoritative voices in the world on authenticating images, video, audio and information itself.
Hany is an applied mathematician and computer scientist. He spent twenty years at Dartmouth, eight at UC Berkeley, and is now back at Dartmouth. For twenty five years he has developed the techniques used by media outlets, law enforcement and courts of law to determine what is real and what is not. He has advised regulators, worked on countering online terrorism and child sexual abuse material, and been asking "who benefits" since long before it was fashionable.
Filmed in our new London studio, this conversation cuts through more of the current AI and social media narrative in ninety minutes than most think pieces do in a year. Hany's argument is direct. If you are getting the majority of your news and information from social media, you should stop.
Topics covered:
The 10 to 1 ratio of AI fake content to real reporting in every major world event
Why disinformation is cheap and information is expensive
The Silicon Valley trillionaire race and what it means for you
The Grok example - Elon Musk hard-coding his own AI to protect his reputation
Why LLMs may be worse than social media for consolidating ideas
The Palo Alto paradox - billionaires banning screens from their own kids' classrooms
The airplane question - would you get on it if the engineers did not understand it
Hany's practical advice on deleting apps, using AI properly, and rebalancing life

This week on GAEA Talks, Graeme Scott sits down with Rob Garlick, former Head of Innovation Research at Citi and author of AI: Anarchy or Abundance? Why the Future of Work Needs Pro-Human Leaders. Rob spent 28 years at Citi and was at the front row for the Netscape and Google IPOs.
Filmed in our new London studio, this is one of the most economically grounded AI conversations on the show. Rob's argument is that this wave is fundamentally different for three reasons - pace, scope and nature of change - and that the biggest economic trade in history is now being made in front of us. People are the largest line item on most corporate P&Ls, and AI now offers a way to substitute them. That is great news for shareholders. It is not automatically great news for people.
Topics covered:
The three reasons this wave is different: pace, scope and nature of change
Milton Friedman's 1970 shareholder primacy essay
Charlie Munger on incentives
The Engels pause of the Industrial Revolution
Humanoid robots with a payback period of under ten weeks
OpenAI agents beating experts 70% of the time at 11x speed and 99% lower cost
The economic singularity and the risk of wage collapse
Universal Contribution Income (UCI) as an alternative to UBI
Rob's BEST framework starting with better jobs
Einstein and Bertrand Russell - "remember your humanity and forget the rest"

This week on GAEA Talks, Graeme Scott sits down with Natalie Monbiot, one of the earliest and most thoughtful voices in the world on digital twins, AI avatars and human-centred AI. Natalie joined the podcast from the US.
Natalie left the corporate world in 2019 to join Hour One as one of its earliest team members, and helped turn AI-generated video from a category dominated by deepfake pornography into a legitimate enterprise business, with customers including Johnson & Johnson and Berlitz. She now collaborates with AE Studio on AI alignment and safety, who recently co-published a paper with Anthropic on removing dangerous knowledge at the training stage.
The title of this episode comes from a question Natalie borrows from the Cosmos Institute. When the AI is not in the room, are you smarter and more capable, or are you diminished? If it is the second, something has gone wrong.
Topics covered:
Are you smarter and more capable when AI is not in the room
Why the digital twin is a healthier mental model than the answer engine
The cognitive atrophy risk of outsourcing your thinking to an LLM
The shift from AI as answer engine to AI as action engine
Yoshua Bengio on stripping AI of any intrinsic personality
John Vervaeke's "relevance realisation" as a distinctly human capability
Sovereign personal AI - your data, your model, on your device
The Echo Studio and Anthropic paper on removing dangerous knowledge at training
Family-safe AI as a real consumer category
Natalie's three prescriptions for using AI in a way that makes you stronger

This week on GAEA Talks, Graeme Scott sits down with Connor Leahy, US Executive Director of Control AI. Connor is one of the most technically credible and politically clear voices in the world on the risks of superintelligence.
Connor has been in AI his whole career. He co-founded EleutherAI, which built some of the first large-scale open-source language models. He then founded Conjecture, a research company focused on controllable and understandable AI systems, before winding it down to move to Washington, D.C. and lead the US work of Control AI.
Filmed in the new London studio, this conversation cuts through more of the current AI narrative in ninety minutes than most policy papers do in a hundred pages. Connor's argument is that superintelligence is not a technical problem, it is a political one. He walks through why modern AI is grown rather than written, why reinforcement learning produces optimising sociopaths by default, why the labs are selling a political product not an economic one, and why "there are no adults in the room" is the single most important thing every leader must understand.
Topics covered:
Why superintelligence is a political problem, not a technical one
Neural networks in plain English - grown, not written
Why we only understand about 3% of what happens inside modern AI
Why reinforcement learning produces optimising sociopaths by default
Why the labs are selling a political product, not an economic one
The leaded gasoline parallel for AI regulation
"Saliency control" as the primary method of mind control
Connor's three prescriptions - write things down, find community, engage policy

Forrester technology and innovation forums (Austin, London, New York). Promo code "GAEATECH" for 10% off. Visit https://forrester.com/events/#tech to book your place.This week on GAEA Talks, Graeme Scott sits down with Dr Aaron Balick, psychotherapist, cultural theorist and author of The Psychodynamics of Social Networking. Aaron is one of the very few people in the world who thinks about AI and the human mind with equal fluency, and this is one of the most useful and quietly unsettling conversations we have recorded.Aaron grew up in Delaware, studied at the University of Colorado Boulder, moved to the UK for a master's in continental philosophy, and then trained as a psychotherapist. His clinical work sits in the very traditional two chairs and a room. His writing and public work sit at the other end of the spectrum, tracing how technology now mediates who we are, how we relate to each other and what we absorb without noticing. In 2013 he published one of the first serious books applying psychoanalytic thinking to social media. He is now revising it for a new edition, and the story has become far bigger than social media.Filmed in our new London studio.The premise Graeme opens with is that fear has been the single most weaponised feature of modern technology, and if we can see that clearly we can start taking back some agency. Aaron picks it up from there and does not let go for the next hour. He walks through Dunbar's number and why we are being asked to relate at a scale humans were never built for. He explains Melanie Klein's paranoid-schizoid position and why fear literally regresses us into black-and-white thinking, making it easier to be manipulated, polarised and lonely. He revisits Cambridge Analytica and the NRA anxiety-targeting example as case studies in how personality data is used to shift intention, not just spend. He introduces his term "digital impingement". He walks through what AI companions are actually doing to attachment. He argues, sharply, that many of these systems are psychopathic in the strict clinical sense of performing empathy without having any. And he closes on the sovereign psyche - the idea that everything Graeme talks about in Sovereign AI has a mirror inside every one of us that we now need to defend and rebuild.