"The Lights Are Off." A Cambridge AI Scientist on What We Are Actually Building.
Episode Title: "The Lights Are Off." A Cambridge AI Scientist on What We Are Actually Building.
Direct Answer: Spencer Kelly, Cambridge computer scientist, BBC Click presenter for 20 years, and AI keynote speaker, argues that the most dangerous assumption in AI is that there is something behind the words it generates. In this episode of The TechDental Podcast, Spencer draws a precise and unsettling distinction between intelligence and consciousness, explains why AI should have been called Applied Statistics, reflects on the Spotify prediction he got spectacularly wrong, and asks whether this technology wave is genuinely different from every one that came before it. Essential listening for anyone building, deploying, or investing in AI in healthcare right now.
What is this episode about?
Spencer Kelly studied artificial intelligence at Cambridge University in the 1990s, before it entered mainstream academic or public discourse. For two decades he presented BBC Click, the corporation's flagship technology programme, reporting from NASA, the Large Hadron Collider, and communities across Kenya, Korea, and India. He left the BBC in March 2025 when Click ended and now delivers keynotes on AI, works as a media coach, and makes music using AI tools.
His central argument is deceptively simple and commercially important. Intelligence and consciousness are not the same thing. AI has extraordinary intelligence by any useful definition of the term. It operates in novel situations, synthesises patterns across vast datasets, and produces outputs that outperform human experts across a growing range of tasks. But there is no awareness behind any of it. No judgement. No experience. Just extraordinarily sophisticated mathematics running in complete darkness.
That distinction, between what AI can do and what it understands about what it is doing, is the most commercially important question anyone deploying AI in a healthcare environment can ask right now. And it is the question most AI conversations avoid entirely.
What you will learn from this episode:
* Why AI should have been called Applied Statistics and what that framing changes about every decision you are about to hand to it
* The distinction between intelligence and consciousness that most people building AI products have never properly examined
* What Spencer got spectacularly wrong about Spotify and what the logic of VC-funded disruption actually teaches you about how technology transforms industries
* Whether this wave of AI is genuinely different from every previous automation cycle and why Spencer holds that question with genuine uncertainty rather than confident prediction
* What 20 years of watching every technology cycle teaches you when the biggest one of all finally arrives
* Why the country with the healthiest relationship with AI right now is Australia
* What AI in general, not just generative AI, will do to healthcare, disease detection, and the patterns no human researcher could ever find alone
Key quotes from this episode:
"There is nothing behind those words. The lights are off. There is no experience. There is no judgement. There is no critical thinking. There is just nothing." — Spencer Kelly
"Half of me thinks surely this is going to be the same. But then the other half thinks, this time it does feel different. Because we have always said the machines can do the dangerous, physical, heavy lifting. But we will still be making the decisions. And AI is creeping into that now." — Spencer Kelly
"He looked at me like I was an idiot. Which I clearly was. I hadn't understood the power of disruption. As long as you have enough money to disrupt and end the previous model, you can do what you want." — Spencer Kelly on Daniel Ek and Spotify
"AI in general is going to improve our lives, but it needs to be used by the right people for the right reasons and not just for profit." — Spencer Kelly
About Spencer Kelly:
Spencer Kelly is a Cambridge computer scientist, technology broadcaster, and AI keynote speaker. He presented BBC Click for 20 years, making him one of the most trusted technology communicators in the UK. He holds a degree in computer science from Cambridge University where his final year dissertation used AI genetic algorithms to optimise the London Underground route. He now delivers keynotes on AI for business audiences globally and works as a media and presentation coach.
Website: https://www.spencerkelly.com [https://www.spencerkelly.com] LinkedIn: https://www.linkedin.com/in/spencer-kelly-400a512/ [https://www.linkedin.com/in/spencer-kelly-400a512/] Instagram: @spenleykelly Twitter: @spenley YouTube: @spencer0kelly
Outro music:
The outro music for this episode is by CLAWS, a band featuring Spencer Kelly's daughter. Used with permission. This is a real-world example of human-AI co-creation: a father and daughter making music together, augmented by AI tools. Spencer discusses this in the episode.
Listen to CLAWS on Spotify: https://open.spotify.com/track/1BoKFIPMePNHMtvY4CKihF [https://open.spotify.com/track/1BoKFIPMePNHMtvY4CKihF] Watch on YouTube: https://youtu.be/7iOo3bWWnQA [https://youtu.be/7iOo3bWWnQA]
About TechDental:
TechDental is a strategic intelligence platform for founders, executives, operators and investors shaping the future of dentistry. Independent analysis on AI, operating models and capital strategy in UK dentistry.
Website: www.techdental.com [http://www.techdental.com] Email: info@techdental.com [info@techdental.com] Host: Dr. Randeep Singh Gill LinkedIn: https://www.linkedin.com/in/drrandeep/ [https://www.linkedin.com/in/drrandeep/] Newsletter: Subscribe on LinkedIn, TechDental Analysis
Subscribe and follow:
Apple Podcasts: https://bit.ly/41pKL9b
Spotify: https://bit.ly/41UsqRO
YouTube: https://bit.ly/3JSfl5c
TechDental publishes independent analysis on AI, operating models and capital strategy in dentistry. For leaders navigating scale or structural change: www.techdental.com [http://www.techdental.com]. Copyright Dr. Randeep Singh Gill and RIG Enterprises Limited (Company No. 11223423) 2026.