Introducing Knowing Machines, a podcast companion to the Knowing Machines project. In this podcast, we're going to look at the data that's used to train artificial intelligence. The building blocks of these systems offer us a powerful way to understand how these systems see the world, how they interpret it, as well as what they don't see, and what they might be getting wrong. So we're here to discuss what that material is and why it matters. The season brings together computer scientists, engineers, social scientists, humanists, and also artists and journalists, a whole range of different interdisciplinary stakeholders in the same room to demystify this moment in artificial intelligence. Episodes will drop weekly. We look forward to have you listen with us.
Kate Crawford:
This is Knowing Machines.
I'm Kate Crawford, and I'm one of the many voices that you'll be hearing over the next several weeks.
So we're in the middle of a world wide experiment with some really high stakes. Artificial intelligence reached an inflection point in 2023. This is the year when systems are being built into almost every industry you can name. It's like the everything, everywhere, all at once for generative AI all across the planet. And many people are understandably drawn to thinking about AI as being almost magical. You know, we type a question into a text box and we get this immediate answer. We can summon a picture of literally anything with just a few words. But there's a lot more going on behind that magic show, and there's a lot more at stake. People are rarely getting access to the deepest stories of what's going on, and this is definitely the moment for it, because we are all part of this world of generative AI, whether we like it or not. Our photographs, our online messages, our videos, every digital trace that we've left online is likely being used to train AI systems right now.
Tom Cruise Deepfake:
I'm going to show you some magic.
Barack Obama Deepfake:
We're entering an era in which our enemies can make it look like anyone is saying anything at any point in time, even if they would never say those things.
Morgan Freeman Deepfake:
I am not Morgan Freeman. And what you see is not real.
Kate Crawford:
In this podcast, we're going to look at the data that's used to train artificial intelligence. The building blocks of these systems, if you will. They offer us a powerful way to understand how these systems see the world, how they interpret it, as well as what they don't see, and what they might be getting wrong. So we're here to discuss what that material is and why it matters.
We've brought together computer scientists, engineers, social scientists, humanists, and also artists and journalists, a whole range of different interdisciplinary stakeholders in the same room to demystify this moment in artificial intelligence. It's been a rare privilege to take some time to think deeply about the longer term consequences of what is being built today.
We're so glad you're here listening with us.
The first episode will drop on Monday, October 30th, and it'll be available at knowingmachines.org or wherever you get your podcasts.