MIT Photographer Reveals the One Thing AI Can Never Replicate

MIT Photographer Reveals the One Thing AI Can Never Replicate

An MIT photographer discusses the unique human curiosity and creativity in photography that AI cannot replicate, using ferrofluid imaging as an example of scientific documentation versus illustration.

MIT Photographer explains one thing AI can never do. | Transcript:

So, I'm curious when you first discovered photography or when you were first wowed by something you caught through a photograph and like the curiosity that was sparked? It was always about my curiosity. Why is this happening? What is it about this thing that is showing this and that? And you have and that's why I'm so excited about speaking with the researchers because they explain it to me and they you know, sometimes it's over my head and I tell them that's what you know, at MIT you never fake what you don't know. Yes, I've learned that. And the best part of what I do is getting the information about the science that I'm imaging. So, for example, at the very beginning when I first started, this was for a book that

George Whitesides and I decided to do. A ferrofluid is oil that has suspended iron particles. Okay. And so, it responds to magnetic fields. So, this is a drop. It's about 2 cm across. I dropped it on a glass slide. Under the glass slide is a yellow post-it. Very high-tech. And under the yellow post-it are seven circular magnets. And what you're seeing is the iron particles responding to the magnetic field. Yeah. And we explain it in the book. And then I also held a green card over it. It was highly reflective, so I decided to add some color which didn't change anything really.

Yeah. And then you see the shapes that are forming. And if you really look carefully, Yes. you can see the sign of a very bad photographer because I'm in the picture. You are? If you look carefully, you can see a lens that's doubling up and that's my head of hair. But that was the beginning of realizing, oh boy, I could have fun with this. You can have fun. And what's very interesting is that this particular image got a lot of attention. This is again many years ago. People are drawn to it. I don't understand it, but it's unusual. It looks like a flower. like a flower, yeah. And so, and then here's the thing. They want to say, what

is this? That's what I'm trying to do. And then that leads them into the science. Precisely. Yeah. That's so interesting. Do you want to use this image to sort of distinguish for us um documentation versus illustration and how manipulation sort of intersects with these two elements? Yeah. This is in fact I haven't thought about this actually. This is a documentation. I'm not adding any Well, yes, I added some color, but if we're only looking at the way the particles are responding to the magnetic field, I am not manipulating that at all. It's it's this is exactly what this is about. There is a very

serious issue about how far can we go and maybe we'll talk about that later in manipulating an image. I that's what I'm very careful about as a science photographer. Sometimes I use images to explain it to explain something. Okay. For example, I have a number of covers that I can't draw for beans, but what I can do is I can make pictures. And I take pieces of pictures and put it together to create an image that doesn't exist, but it's explaining the science. So, that would be an illustration. Exactly. That is a very big difference between an illustration and documentation. And thank you for concentrating on that because that is the issue that we're going to

have to deal with when we look at AI.

More Entertainment Transcript