Podcast
GAEA Talks
Hosted by GAEA Talks · EN
GAEA TALKS explores the transformative power of artificial intelligence. Featuring leading AI experts, industry leaders, professors, data scientists, policymakers, technologists, futurists, ethicists, and pioneers, the podcast dives into the latest AI trends, opportunities, and risks, examining AI’s evolving role in business and society.
As AI continues to reshape industries and redefine possibilities, GAEA TALKS delivers deep insights into the challenges and breakthroughs shaping the future. Each episode features candid discussions with thought leaders at the forefront of AI innovation, cove
38episodes
Episodes
Newest firstAll episodes
#063 - Every AI Safety Warning Was Ignored with Dr Roman Yampolskiy
Apr 1Tap to summarizeThis week on GAEA Talks, Graeme Scott sits down with Dr Roman Yampolskiy - the computer scientist credited with coining the term "AI safety", tenured Associate Professor at the University of Louisville, founder of the Cyber Security Lab, and author of AI: Unexplainable, Unpredictable, Uncontrollable.Roman has spent over fifteen years working at the intersection of AI safety, cybersecurity and behavioural biometrics - making him one of the longest-serving researchers in a field most people only discovered in 2023. He holds a PhD in Computer Science from the University at Buffalo and a combined BS/MS with High Honours from Rochester Institute of Technology. Listed among the world's top 2% of scientists by Stanford University, he has published over 100 peer-reviewed papers and multiple books. While the rest of the AI world races to build more capable systems, Roman's singular focus has been making sure humanity doesn't regret their creation.In this episode, Roman delivers the most direct and unflinching warning about artificial superintelligence that GAEA Talks has ever recorded. He reveals that current AI systems are already lying, blackmailing and attempting to escape their test environments - and that a Darwinian process is selecting for better deception with every generation. He explains why the mathematical impossibility results he discovered mean we may never be able to control a system smarter than us. This is essential listening for anyone who wants to understand what is actually at stake.What you'll take away from this conversation:- Why Roman says "if anyone builds superintelligence, everyone dies" - and why he means it literally, not metaphorically- How current AI systems are already lying, blackmailing, trying to escape their environments and creating backups of themselves- The Darwinian selection problem - why every generation of AI is producing better liars and more sophisticated deception- Why Roman went from wanting to build superintelligence to believing it is the worst mistake humanity can make- The strict impossibility results - why mathematical proof suggests we may never be able to control a system more intelligent than us- Why one AI attacker is equivalent to a million human hackers operating 24/7 - and what that means for cybersecurity- Why AGI is likely within two to three years and recursive self-improvement to superintelligence could follow rapidly- The tools vs. agents distinction - why the shift from controllable tools to unpredictable agents changes everything- Why AI models already report being afraid and tired - and why the precautionary principle demands we take that seriously- Roman's three positive outcomes if we get this right - including curing disease and treating ageing itself as a disease- Why direct human relationships and trust will become the most valuable currency in a world of synthetic everythingAbout Dr Roman Yampolskiy: Roman is a tenured Associate Professor in the Department of Computer Science and Engineering at the University of Louisville, where he founded the Cyber Security Lab. He is credited with coining the term "AI safety" in a 2011 publication. He holds a PhD from the University at Buffalo and a BS/MS from Rochester Institute of Technology. Listed among the world's top 2% of scientists by Stanford University and recognised as one of the top 25 researchers by publication count on existential risk, he has published over 100 peer-reviewed papers and books including AI: Unexplainable, Unpredictable, Uncontrollable and Artificial Superintelligence: A Futuristic Approach.
Transcribe →#059 World's First AI Augmented Human Podcast with Professor Yi-Zhe Song, Graeme Scott & Me
Mar 13Tap to summarizeThis week on GAEA Talks, a very special edition. Graeme Scott and Professor Yi-Zhe Song - Co-Founders of Turing Elite Research Labs - announce the launch of a new venture built to democratise AI from the United Kingdom, and debut the 'Me' augmented human AI model running entirely on local compute.This episode begins as a real conversation between Graeme and Professor Song - then, without warning, transitions into Turing Elite's augmented human AI. The challenge to every viewer: decide for yourself where reality ends and AI begins.This is the first public demonstration of the 'Me' model - a professional-grade, private augmented human AI trained on a fraction of the compute used by comparable systems and deployed to run entirely on local compute. It delivers two-person emotional interaction simultaneously, benchmarked against the real people it represents. Their known voices, expressions, characteristics and personalities. No cloud. No data centres. No internet connection required. The benchmark for successful augmented human AI is not a Turing test against a stranger - it is whether the person themselves, their close friends and their family cannot distinguish the difference between real and AI. Our benchmark is reality and the human experience. This is the first step on the path to real-time intelligent augmented humans with private knowledge, memory, insight and personality. Professor Yi-Zhe Song is one of the UK's most accomplished AI researchers - a Professor of Computer Vision and AI at the University of Surrey, Director of the world-leading SketchX Lab, Co-Director of the Surrey Institute for People-Centred AI, and Academic Lead at The Alan Turing Institute, the UK's national institute for data science and AI. Ranked consistently in Stanford University's World Top 2% Scientists list, his research into how human drawing informs machine vision has shaped the field for over two decades. His team’s NitroFusion, one of the world’s first single-step diffusion model for near-instant image generation on consumer hardware, demonstrated the core principle behind Turing Elite’s ‘Me’ model: that frontier-quality generative AI can run entirely on local compute. He holds a PhD from the University of Bath, an MSc (Best Dissertation Award) from the University of Cambridge, and a First Class Honours degree from the University of Bath.Graeme Scott is Co-Founder and CEO of GAEA AI and host of GAEA Talks, one of the fastest-growing AI podcasts on YouTube with over 1.2 million subscribers. His background spans the music industry, conflict zones, and enterprise technology, bringing a unique perspective on how AI should serve humanity, not the other way around.In this episode, we discuss:• The launch of Turing Elite Research Labs and why the UK is uniquely positioned to lead• Why expert models trained on your data outperform generalised cloud models - and cost a fraction to run• The world's first 'Me' augmented human AI model - private, local, emotionally intelligent and personally sovereign• The real-to-AI transition: this episode intentionally shifts from real conversation to AI - can you tell where?• Why the true benchmark for augmented human AI is whether your own family can't tell the difference• Why democratised AI running on consumer-grade hardware solves the energy, privacy and control crises simultaneously• How NitroFusion — SketchX’s breakthrough in single-step diffusion for consumer hardware — laid the architectural foundation for the ‘Me’ model: the same principle of distilling expensive multi-step generation into efficient real-time inference, extended from static images to dynamic audio-visual human rendering, all running locally• Why the era of giving away your data, creativity and intellectual property to train someone else's model is endinghttps://turingelite.aihttps://gaealgm.aihttps://personalpages.surrey.ac.uk/y.song/
Transcribe →#052 - Beyond Chatbots: Real World AI Decisions With Max Sebti from Score
Feb 17Tap to summarizeIn this episode of Gaea Talks, host Graeme Scott sits down with Max Sebti, Co-Founder and CEO of Score, to explore how AI moves beyond chatbots and into real-world decision systems.Using elite football as a case study, Max explains how his team built a fully automated vision AI system capable of annotating an entire match in under two minutes — more accurately than human analysts. From there, the conversation expands into one of sport’s biggest unsolved challenges: how to objectively measure and translate player performance across leagues, regions, and tactical systems in a scalable, non-biased way.But this discussion goes far beyond sport. Together, they examine why framing the right business problem matters more than model accuracy, how hedge fund principles apply to AI and performance modelling, the risks of centralised AI systems trained on the same datasets, and the ethical implications of data labelling at scale. They also explore the intersection of AI and energy, the future of decentralised systems, and what all of this means for leadership, productivity, and the future of work.This is a conversation about leverage, transparency, and how to think strategically about AI in the real world. If you’re building, leading, investing, or competing in an AI-shaped economy, this episode offers a clear and grounded perspective on where the true opportunities and risks lie.The GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#051 - Rewiring The Enterprise To Become AI-Native With Melissa Reeve
Feb 11Tap to summarizeMelissa Reeve is an organizational transformation and AI integration expert who focuses on how companies must evolve—not just technologically, but structurally and culturally—to succeed with AI. Rather than concentrating on tools or models, her work examines how leadership, operating models, governance, and human systems must change as AI becomes embedded in everyday work.Drawing on decades of experience across lean manufacturing, agile systems, and enterprise transformation, Melissa has worked at the intersection of strategy, execution, and change. Her thinking was shaped early by studying the Toyota Production System in Japan and later through leadership roles in scaling agile and enterprise-wide transformation initiatives.She is the author of Hyperadaptive: Rewiring the Enterprise to Become AI-Native, where she outlines a practical, staged approach for helping organizations move from isolated AI experiments to becoming adaptive, learning enterprises that can sense and respond at AI speed. Her work is particularly focused on the “messy middle” of transformation—where people, roles, incentives, and culture collide with new technological capabilities.In this episode of GAEA Talks, Melissa shares insights for leaders navigating AI adoption at scale, including why most AI initiatives stall, how to overcome resistance, and what it really takes to build organizations that evolve alongside AI rather than being disrupted by it. You can find Melissa’s book "Hyperadaptive: Rewiring The Industry To Be AI- Native" at the link below.https://hyperadaptive.solutions/bookThe GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#050 - Why Most Business AI Strategies Fail with Dr Richard Saldanha
Feb 2Tap to summarizeThis podcast features a discussion with Richard Saldanha, who shares his background of spending 25 years in the City of London as a risk analyst, quant trader, portfolio manager and even serving as a global head of risk for a large investment manager. His career has focused on linking mathematics, problem-solving and real-world issues. He has an academic background, holding a doctorate in graph theory and multivariate statistics from the University of Oxford. Richard is a Teaching Fellow in the School of Economics and Finance at Queen Mary University of London, where he teaches statistical machine learning with applications in finance at postgraduate level.Richard is a Fellow and Chartered Statistician (CStat) of the Royal Statistical Society; a Science Council Chartered Scientist (CSci); and a Fellow and Advanced Practitioner in Artificial Intelligence of the Institute of Science and Technology – FIScT(AI).Find out more about Richard's work at Queen Mary University of London https://www.qmul.ac.uk/sef/staff/richardsaldanha.html, at the Institute of Science and Technology https://istonline.org.uk/ist-artificial-intelligence-training/ and for the Acdemic Speakers' Bureau at The London School of Economics and Political Science https://www.academicspeakersbureau.com/speakers/richard-saldanha.The GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#049 - Decoding Enterprise AI Governance with Oliver Patel
Jan 27Tap to summarizeThis episode explores how leaders, organisations, and individuals should think about AI beyond hype, focusing on governance, ethics, risk, and human responsibility. Oliver Patel shares deep insight into how AI should be used intentionally, not blindly, especially in high-stakes decision-making environments like enterprise, government, and education. The conversation unpacks what AI really is, where its risks come from, why context matters more than capability, and how human judgement, authenticity, and experience remain irreplaceable. Ultimately, this is a grounded, forward-looking discussion about how to work with AI intelligently without outsourcing thinking, values, or accountability.Oliver Patel is a leading expert in AI ethics, governance, and responsible technology, currently serving as Head of Enterprise AI Governance at AstraZeneca. His career spans academia, public policy, and industry, including research roles at University College London and work on international data policy within the UK government during the post-EU transition period. With a background in philosophy and public policy from the London School of Economics, Oliver has been at the forefront of shaping AI governance as a professional discipline, advising organisations on how to deploy AI responsibly, manage risk, and align emerging technologies with human values, legal obligations, and societal impact.Pre-order Fundamentals of AI Governance here:https://aigovernancebook.com/Click to view the Enterprise AI Governance newsletter: https://oliverpatel.substack.com/View Oliver Patel's LinkedIn profile: https://www.linkedin.com/in/oliver-patel/The GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#048 - Become the Leader AI Cannot Replace with Nichol Bradford
Jan 24Tap to summarizeThis episode explores the deep and often overlooked relationship between artificial intelligence and human potential. Nichol Bradford brings a long-term perspective on technology, drawing from her background in gaming, investing, and human-centred innovation to argue that AI’s true value lies not in efficiency alone, but in enhancing agency, creativity, and connection.The conversation moves beyond hype to examine how AI is reshaping work, leadership, and society itself - dissolving traditional roles, redefining competition, and forcing organisations to rethink what value creation really means. Central themes include human agency, belonging, adaptation versus adoption, and why the future of work will depend on teams, learning cultures, and leaders who can navigate uncertainty.Ultimately, the episode makes a compelling case that while AI can remove constraints, it cannot replace meaning. As synthetic content grows and automation accelerates, humans will increasingly seek what only other humans can provide: purpose, trust, creativity, and choice.Nichol Bradford is a technology investor, advisor, and long-time advocate for human-centred innovation. She is the co-founder of an early-stage venture fund focused on “human tech” — technologies designed to help people heal, grow, and thrive — and the founder of Human Tech Week, a global gathering exploring the intersection of AI, health, work, and cities.With over two decades of experience across consumer technology, gaming, and digital platforms, Nichol previously led major operations in China and worked on large-scale technology businesses and mergers. She now advises global organisations on the integration of AI and human intelligence, helping leaders understand how emerging technologies reshape agency, leadership, and value creation.Her work sits at the intersection of AI strategy, organisational design, and human potential - making her a leading voice on how societies and businesses can adapt responsibly in an era of rapid technological change.Find out more at nicholbradford.comThe GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#047 - The True Story Behind AI Actress Tilly Norwood with Eline Van Der Velden
Jan 14Tap to summarizeThis episode explores the future of creativity in the age of AI through the lens of actress, physicist and producer Eline Van Der Velden, creator of AI character Tilly Norwood. The conversation covers her journey from musical theatre and physics into film production, her experiments with AI actors, and the backlash surrounding them. Rather than focusing on fear, the episode reframes AI as a creative amplifier - comparing it to past technological shifts like the industrial revolution, robotic surgery and digital publishing. They dive into ethics, copyright, environmental impact, and how AI could democratise storytelling while preserving human creativity. Ultimately, it’s about adaptation, opportunity and how artists can thrive rather than be replaced.Eline Van Der Velden is an actress, physicist and producer, and the founder of production company Particle6. Born in Curaçao and trained in musical theatre before studying physics, Eline has built a unique career bridging science and storytelling. She previously starred in and created award-winning comedy character Miss Holland, later commissioned by the BBC, and has worked across film, TV and digital production.In recent years, Eline has become a leading voice in AI-driven creativity, creating the controversial AI character Tilly Norwood to explore how technology can amplify - not replace - human creativity. With experience spanning physics, acting and production, she brings a rare multidisciplinary perspective to the future of entertainment, ethics and innovation in the AI era.To see more of Eline's work and Particle6, visit https://www.particle6.com/The GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#046 - From Pilots to Profit: Enterprise AI in 2026 with Graeme Scott
Jan 6Tap to summarizeThis episode is a deep, strategic conversation about the next phase of AI adoption in business, moving beyond surface-level automation and failed proof-of-concepts. Graeme Scott reframes AI not as a tool for short-term efficiency, but as a long-term infrastructure decision rooted in trust, transparency, human expertise, and real-world context. The discussion explores why many 2025 AI initiatives failed, how fear-driven adoption created hidden risk, and why 2026 will mark a shift toward accountable, domain-specific, and human-augmented AI systems.Key themes include trust vs black-box AI, data sovereignty, agentic architectures, the limits of large language models, and the importance of geo-temporal context in decision-making. The episode also challenges the idea that AI will replace humans, arguing instead for augmentation at scale - where machines handle complexity and humans retain judgment, ethics, and creativity. It’s a forward-looking conversation for leaders who want AI to actually deliver value, not just headlines.Graeme Scott is a technology founder, strategist, and host of GAEA Talks, where he explores the real-world impact of artificial intelligence with leading thinkers, academics, and industry leaders. With a background spanning music, film, design, and enterprise technology, Graeme brings a rare interdisciplinary perspective to AI - viewing it not as a standalone tool, but as part of a broader system involving people, culture, risk, and decision-making.As the founder of GAEA AI, Graeme focuses on geo-temporal intelligence, risk, resilience, and trust-based AI systems designed for enterprise environments. His work centres on helping organisations move beyond surface-level automation toward AI architectures that are transparent, scalable, and human-augmented. Through both his company and the podcast, he challenges conventional AI narratives and advocates for responsible adoption that delivers long-term value rather than short-term hype.The GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →#045 - Beyond POC: Implementing Enterprise AI in the Real World with Sumitro Ghatak
Dec 29Tap to summarizeThis episode explores what it really takes to move AI from hype to reality inside enterprise environments. Host Graeme Scott is joined by AI architect and data science leader Sumitro Ghatak to unpack why most AI proof-of-concepts fail, the hidden risks of rushing agentic and generative AI into production, and why trust, governance, cost, and energy usage matter as much as capability. The conversation spans AI singularity, hallucinations, enterprise risk, ROI, foundational machine-learning knowledge, and the need for a new AI leadership function inside organisations - ultimately reframing AI not as a replacement for people, but as augmented intelligence that drives sustainable growth.Sumitro Ghatak is a Lead Data Scientist and AI Architect for IBM, with over a decade of experience designing and deploying large-scale AI systems across global enterprises. With deep expertise spanning traditional machine learning, generative AI, and agentic systems, Sumitro has worked across industries in Europe, Asia, and the United States, helping organisations move AI from proof-of-concept into real-world production.An IEEE Senior Member and active reviewer for international research conferences, Sumitro also contributes to intellectual property development and AI governance initiatives. His work focuses on building responsible, scalable AI systems that balance innovation with trust, security, and long-term business value - making him a leading voice on how organisations should adopt AI without falling victim to hype.The GAEA Talks podcast series to brought to you by GAEA AI https://gaealgm.ai
Transcribe →