Subvolts take DeepMind researchers explain how robots learn physical tasks and address AI safety. Aug 20, 2019 - Developers
- Learn about the challenges and approaches to teaching AI physical tasks and ensuring safety in robotic applications.
- Business
- Understand the practical hurdles and safety considerations for deploying AI and robotics in real-world scenarios.
- Everyday users
- Discover how AI is learning to perform physical tasks and the important questions about its safe use in everyday life.
This episode features researchers from a leading AI lab discussing fundamental challenges in robotics and AI safety.
AI, Robot
Audio streams from Hannah Fry. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
This Google DeepMind podcast episode, hosted by Hannah Fry, explores embodied AI and robotics. Researchers discuss learning tasks from scratch, like robot arms picking up bricks, and the challenges of integrating AI and robotics. The conversation also touches on crucial questions about the safe deployment of AI in the real world, featuring insights from DeepMind researchers.
Imagine teaching a robot to do simple things, like stacking blocks, just like a baby learns. This episode shows how scientists at DeepMind are teaching robots to learn new tasks from scratch. It's about making AI useful and safe in the real world.
This episode offers a grounded look at AI and robotics, moving beyond sci-fi tropes. It's for anyone curious about how AI learns physical tasks and the practical challenges of AI safety. The discussion on embodied AI and learning from scratch connects directly to current frontier lab efforts in developing more capable and safe AI systems that interact with the physical world.
- HostHannah Fry
- GuestJackie KaySoftware engineer, DeepMind
- GuestMurray ShanahanResearch scientist, DeepMind
- GuestVictoria KrakovnaResearch scientist, DeepMind
- GuestRaia HadsellResearch scientist, DeepMind
- GuestJan LeikeResearch scientist, DeepMind
- Episode Title: AI, Robot
- Show: Google DeepMind: The Podcast
- Length: 33 minutes
- Topic Focus: Embodied AI and robot learning
- Key Discussion Area: AI safety in the real world
Forget what sci-fi has told you about superintelligent robots that are uncannily human-like; the reality is more prosaic. Inside DeepMind’s robotics laboratory, Hannah explores what researchers call ‘embodied AI’: robot arms that are learning tasks like picking up plastic bricks, which humans find comparatively easy. Discover the cutting-edge challenges of bringing AI and robotics together, and learning from scratch how to perform tasks. She also explores some of the key questions about using AI safely in the real world.
If you have a question or feedback on the series, message us on Twitter (@DeepMind using the hashtag #DMpodcast) or email us at podcast@deepmind.com.
Further reading:
- Blogs on AI safety and further resources from Victoria Krakovna
- The Future of Life Institute: The risks and benefits of AI
- The Wall Street Journal: Protecting Against AI’s Existential Threat
- TED Talks: Max Tegmark - How to get empowered, not overpowered, by AI
- Royal Society lecture series sponsored by DeepMind: You & AI
- Nick Bostrom: Superintelligence: Paths, Dangers and Strategies (book)
- OpenAI: Learning from Human Preferences
- DeepMind blog: Learning from human preferences
- DeepMind blog: Learning by playing - how robots can tidy up after themselves
- DeepMind blog: AI safety
Interviewees: Software engineer Jackie Kay and research scientists Murray Shanahan, Victoria Krakovna, Raia Hadsell and Jan Leike.
Credits:
Presenter: Hannah Fry
Editor: David Prest
Senior Producer: Louisa Field
Producers: Amy Racs, Dan Hardoon
Binaural Sound: Lucinda Mason-Brown
Music composition: Eleni Shaw (with help from Sander Dieleman and WaveNet)
Commissioned by DeepMind
Please leave us a review on Spotify or Apple Podcasts if you enjoyed this episode. We always want to hear from our audience whether that's in the form of feedback, new idea or a guest recommendation!
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
- embodied AI
- Murray Shanahan, Victoria Krakovna, Raia Hadsell, Jan Leike
- AI that learns by interacting with the physical world, often through robots.
44 minCan AI help us better predict the weather?
WithPeter Battaglia
45 minThe mathematics of AI uncertainty
WithZoubin Ghahramani
53 minSame guests, other shows
41 minConsciousness, Reasoning and the Philosophy of AI with Murray Shanahan
WithMurray Shanahan
36 minMe, myself and AI
WithDemis Hassabis, Raia Hadsell, Karl Tuyls, Zach Gleicher
34 minLet's get physical
WithRaia Hadsell, Viorica Patraucean, Jan Humplik, Akhil Raju
26 minTowards the future
WithKoray Kavukcuoglu, Trevor Back, Raia Hadsell, Murray Shanahan
59 minQuestions people ask
- What is embodied AI?
- Embodied AI refers to AI systems, often in robotic form, that learn tasks by interacting with the physical world, like robot arms learning to pick up objects.
- What are the key challenges discussed?
- The episode covers the cutting-edge challenges of bringing AI and robotics together, focusing on how robots learn tasks from scratch and the critical questions surrounding AI safety in real-world applications.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Hannah Fry.