Subvolts take AI needs self-doubt. Learn how math builds machine humility. Aug 26, 2026 - Developers
- Understand how incorporating uncertainty can lead to more robust and reliable AI models.
- Business
- Learn about a key research area that could improve AI's trustworthiness and performance in real-world applications.
Features a leading researcher discussing a core AI challenge, offering insights into future AI development.
The mathematics of AI uncertainty
Audio streams from Hannah Fry. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Google DeepMind's podcast features Zoubin Ghahramani discussing the mathematics of AI uncertainty. Ghahramani, a professor and VP at Google DeepMind, has researched uncertainty in AI for 30 years. The episode explores how giving machines a sense of their limitations could be crucial for improving AI, contrasting with current AI's confident, often incorrect, answers.
Imagine AI always sounding sure of itself, even when wrong. This episode explains how mathematicians are teaching AI to understand its own limits, like humans do. This could make AI more reliable and trustworthy.
This episode delves into a foundational aspect of AI development: uncertainty. It's for researchers and AI enthusiasts interested in moving beyond current large language models' limitations. Ghahramani's work on Bayesian thinking and uncertainty directly relates to ongoing efforts at Google DeepMind and other frontier labs to build more robust and reliable AI systems, potentially addressing issues like AI hallucination and overconfidence.
- HostHannah Fry
- GuestZoubin GhahramaniProfessor and VP of Research, Google DeepMind
- Guest: Zoubin Ghahramani
- Title: The mathematics of AI uncertainty
- Show: Google DeepMind: The Podcast
- Length: 45 minutes
- Ghahramani's research focus: Mathematics of uncertainty in AI for 30 years
Now, if you ask an AI a question, it will usually give you an absolute answer with unwavering authority, even if that answer turns out to be wrong. In fact, today's AI seems to be missing a fundamental human trait: self-doubt. Long before the current wave of large language models, one academic researcher was trying to give machines a sense of their own limitations. Zoubin Ghahramani has spent the last 30 years pioneering a type of intelligence built on the mathematics of uncertainty. Today, as a professor at Cambridge and VP of Research at Google DeepMind, Zoubin finds himself at the heart of another interesting debate: will improving machine uncertainty be one of the missing pieces to ever improving AI?
Timecodes:
- 00:00 Introduction
- 01:06 The role of uncertainty
- 07:45 Correctness vs confidence
- 09:40 Historical perspectives
- 16:10 Bayesian thinking in AI
- 26:30 Uncertainty in the real world
- 36:42 Future research and AGI
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.
- Bayesian thinking
- Zoubin Ghahramani
- A way of thinking about probability and updating beliefs as new information becomes available.
- AGI
- Zoubin Ghahramani
- Artificial General Intelligence, a hypothetical AI with human-like cognitive abilities.
44 minCan AI help us better predict the weather?
WithPeter Battaglia
53 min
54 min
1 h 06World Models and the Future of Spatial AI with Justin Johnson - #775
WithJustin Johnson
Questions people ask
- What is the main problem with current AI answers?
- Current AI often gives absolute answers with unwavering authority, even when those answers are wrong. They lack a human trait of self-doubt.
- What is Zoubin Ghahramani's area of expertise?
- Ghahramani pioneered intelligence built on the mathematics of uncertainty and has researched this for 30 years.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Hannah Fry.