Science Friday - AI Helps Find Ancient Artifacts In The Great Lakes | An Artist Combines Indigenous Textiles With Modern Tech

Episode Date: January 25, 2024

Researchers in Michigan modeled a prehistoric land bridge and used AI to predict where caribou–and humans–might have traveled along it. Also, artist Sarah Rosalena uses Indigenous weaving, ceramics, and sculpture practices to create art that challenges tech’s future.Using AI To Help Find Ancient Artifacts In The Great LakesAt the bottom of Lake Huron there’s a ridge that was once above water. It’s called the Alpena Amberley Ridge and goes from northern Michigan to southern Ontario. Nine thousand years ago, people and animals traveled this corridor. But then the lake rose, and signs of life were submerged.Archaeologists were skeptical they’d ever find artifacts from that time. But then John O’Shea, an underwater archaeologist based at the University of Michigan, found something. It was an ancient caribou hunting site. O’Shea realized he needed help finding more. The ridge is about 90 miles long, 9 miles wide and 100 feet underwater.“Underwater research is always like a needle in a haystack,” said O’Shea. “So any clues you can get that help you narrow down and focus … is a real help to us.”That’s where artificial intelligence comes in. He teamed up with computer scientist Bob Reynolds from Wayne State University, one of the premier people creating archaeological simulations. And Reynolds and his students created a simulation with artificially intelligent caribou to help them make predictions.An Artist Combines Indigenous Textiles With Modern TechWhen multidisciplinary artist Sarah Rosalena looks at a loom, she thinks about computer programming. “It’s an extension of your body, being an algorithm,” she says.Rosalena, a Wixárika descendant and assistant professor of art at the University of California Santa Barbara, combines traditional Indigenous craft—weaving, beadmaking, pottery—with new technologies like AI, data visualization, and 3D-printing. And she also works with scientists to make these otherworldly creations come to life. She involved researchers at the NASA Jet Propulsion Lab to make 3D-printed pottery with simulated Martian clay. And she collaborated with the Mount Wilson Observatory to produce intricately beaded tapestries based on early-1900s glass plates captured by the observatory’s telescope, which women mathematicians used to make astronomical calculations.And that’s also a big focus for Rosalena: spotlighting the overlooked contributions women made to computer science and connecting it to how textiles are traditionally thought of as a woman-based craft. When she first started making this kind of art, Rosalena learned that the Jacquard loom—a textile advancement in the 1800s that operated on a binary punch card system which allowed for mass production of intricate designs—inspired computer science pioneer Ada Lovelace when she was developing the first computer program. “[They] have this looped history,” she says. “And when I weave or do beadwork, it’s also recalling that relationship.”But for Rosalena, there is tension and anxiety in her decision to combine new and ancient mediums. “We’re at this point of the technological frontier and that’s actually terrifying for a lot of people, especially for people from my background and my Wixárika background,” she says. “It’s progress for some, but it’s not for all.”Part of Rosalena’s work is anticipating future forms of colonization, especially amid rapid change in our planet’s climate and the rise of AI. “What happens when we bring traditional craft or Indigenous techniques with emerging technology to think about current issues that we are facing? Digital technologies are always chasing after ways that we could simulate our reality, which also produces this way that we could re-envision our reality,” she says.SciFri producer and host of our podcast Universe Of Art D. Peterschmidt sat down with Rosalena to talk about how she approaches her work, why she collaborates with scientists, and how she hopes her art makes people consider today’s technological advancements through an Indigenous lens.Transcripts for each segment will be available after the show airs on sciencefriday.com. Subscribe to this podcast. Plus, to stay updated on all things science, sign up for Science Friday's newsletters.

Researchers in Michigan used AI to recreate a prehistoric land bridge, complete with digital wildlife. She described the caribou looked like they were roller skating because they didn't exactly walk. They say it's Thursday, January 25th. But we know it's really Science Friday. I'm SciFri producer Charles Burquist. In this episode, we'll see how archaeologists are using AI to track the paths of prehistoric caribou to see where artifacts from ancient hunters might be located. But first, a conversation about indigenous art and its intersection with spaceflight.

Here's Ira Flato. The patterns woven in textiles can tell a powerful story, and Sarah Rosalina knows this well. She's a multidisciplinary artist who blends ancient mediums and indigenous knowledge with data and new technology. She's collaborated with NASA JPL, the LA County Museum of Art Tech Lab, and her work is currently featured at the Columbus Museum of Art in Columbus, Ohio, until February 4th. Sci-Fry producer and host of our podcast, Universe of Art, Dee Petersmith, sat down with Rosalina to talk about her collaborations with scientists, space colonization, and how she views technological advancements through an indigenous lens.

Here's Dee. When Sarah Rosalina thinks about the loom, she thinks about computer programming. It's an extension of your body being an algorithm. Ada Lovelace, who wrote the first algorithm design for a computer, said she'd been inspired by the Jakarta loom developed in the 1800s, which used a binary punch card to mass-produce intricate textile designs. And that approach, blending old mediums with new tech, sums up Rosalina's approach to her own art. She's an assistant professor of art at UC Santa Barbara, based in L.A. And she's of Warwicka descent, indigenous people native to what is now parts of Mexico and the southwestern United States. She works in these old art forms, textiles and pottery, but uses AI and data visualization as part of the creative process.

It's a way to process her feelings about how modern society is progressing. We're at this point of the technological frontier, and that's actually terrifying for a lot of people, especially for people from my background and my Warwicka background. We're living in the time of climate change, dispossession, the rise of AI. And I'm always interested in anticipating future forms of colonization. because it's progress for some, but it's not for all. Rosalina, whose fourth generation Warroke Weaver, was taught indigenous textile work in part by her grandmother. It's something that really made her feel sane.

I remember she used to always encouraged me to weave for mental health, but it was also good for exercising your mind. Rosalina later found herself in the Bay Area around the time of the tech boom of the late aughts and learned to code. And there was a lot of interesting people that I met at that time were very similar to me. A lot of WIPOP people working in code. But at the same time, tech startups really started to rise

and displacing a lot of the people that I used to enjoy hanging out with. Frustrated, she moved back to L.A. and rediscovered her love for textile work. I saw so many relationships between, you know, the code that I was writing and actual designs that I was weaving, that they couldn't help but intersect. It was very much like an aha moment. What happens when we bring traditional craft or indigenous techniques

with emerging technology to think about current issues that we are facing. Digital technologies are always chasing after ways that we could simulate our reality, which also produces this way that we could re-envision our reality. And Rosalina doesn't just re-invision reality with herself. She often collaborates with scientists to make her art. It's a big role as an artist to work with scientists and engineers because we see the world differently, and there's a lot of value in that. One of those collaborations was with NASA JPL in Pasadena,

and Rosalina learned that they had a mutual interest, clay. The space agency was experimenting with simulated Martian soil, also called Regolith, to potentially construct livable human habitats on the red planet without having to transport heavy building materials all the way from Earth. So they were doing a lot of research on Regulus, simulant, and clay, to actually build some of the first elements out of basically Adobe, which made me giggle, because, again, it's like how much space colonization is dependent on indigenous knowledge, even on another planet.

Rosalina also teaches coil pot construction at UC Santa Barbara, an indigenous method of making ceramics. That's one of the oldest in the world. Coil pots look like what they sound like. Coils of clay are layered on top of each other until you get your vessel. And she wanted to update that with a techie Martian twist. With the help of NASA engineers, she was able to make her own version of Martian clay based off soil analyses from J.P.H.

She modeled the vessels in the computer and then 3D printed them using her Martian Adobe. The resulting sculptures look both futuristic and ancient. The ribbed, rust-tinged ceramics take a few shapes, the mouth of a black hole, a long cylinder that looks like it's eating itself, a vaguely spherical shape that appears as though it was crushed by the forces of gravity. And Rosalina's passion for pottery even rubbed off on JPL's engineers. I actually made a lot of friends. Some of them got into ceramics at the time, which was also really interesting to have Martian cartographers who were guiding the rovers suddenly be interested in actually the chemical compounds of play. And we would talk for hours on end on making clay, finding native clay in Los Angeles. Located just a few miles away from JPL is the Mount Wilson Observatory, which Rosalina has also partnered with.

It was an important observatory in the early 1900s. Edwin Hubble used a telescope to prove that the universe is expanding, but discoveries like that couldn't have been made without the help of female computers. Women who analyzed the raw data from the telescope and performed complex math that made those discoveries possible. But when I got there, I realized that female computers were mostly cropped and edited out of the history of that observatory. Back then, the images from the telescope were exposed onto glass plates,

which the female computers used to make their calculations. And I found textile was a unique way to approach it because it is a feminist or a female-based craft. So to shed light on these women's work, Rosalina took those plates and digitized them into a lower resolution where each pixel would become a bead on a tapestry, which she then assembled by hand. But not all of these tapestries are neat rectangles. Some of them distort and fray as the beads progress downwards, looking like a starry cosmic jellyfish. Rosalina hopes or art doesn't just serve as a form of protest, but also provides an alternative way of interpreting the world around us, one that places a much larger emphasis on indigenous knowledge. It is very important because a lot of the current crisis that we're facing are a crisis of humanity in many ways.

And I feel like artists really shine that light and also can see the world differently than what a scientist or engineer does. And we can learn quite a bit from one another. For Science Friday, I'm Dee Peter Schmidt. Thanks, Dee. You can check out photos of Rosalina's work at ScienceFriday.com slash textile art. And like I said before, her art is on view at the Columbus Museum of Art in Columbus, Ohio, until February 4th. Artificial intelligence is great at detecting patterns, which means its calculations can help predict the future. But AI can also be used to take a look back into the past. That's exactly what one research team in Michigan is doing, using AI to track the paths of prehistoric caribou.

Why? To see where artifacts from ancient hunters may be located. Joining me to talk about this is my guest, Morgan Springer, editor of the Points North podcast at Interlock in Public Radio in Interlock in Michigan. Welcome to Science Friday. Thank you so much for having me. Help me imagine what we're talking about here before we get to this. the AI caribou. Where is this land bridge that researchers are so interested in? Yeah, so it's at the bottom of Lake Huron, which for listeners that don't know, it's one of the Great Lakes. It's on

the east side of Michigan. And the official name of the land bridge is the Alpina Amberly Ridge. And it goes from northern Michigan to southern Ontario, kind of cutting the lake at a diagonal. And what's the significance of this bridge? Yeah, so what I'm going to say, it's going to sound obvious once I say it, but the Great Lakes didn't always look the way that they do now. If we go back to the ice age, the glaciers are receding. And about 10,000 years ago, lake levels were lower than they are today. So that means land that's now underwater. It was above water then, including this ridge, this land bridge. And it was continuous. It was this causeway where people and animals could move and migrate back and forth and leave artifacts, presumably. And so then the water levels rise,

it comes up, and these artifacts are submerged and remarkably preserved and protected from development. And, you know, archaeologists were skeptical that they'd find anything, that this was going to be an opportunity to find artifacts, but they wanted to look anyway. And let's get into the details of this. Why would someone want to research how animals crossed this long, gone path. Yeah. So John O'Shea, he's an anthropological archaeologist and he's based at the University of Michigan and he wanted to find something. And so basically he came up with an idea for something he thought he could find, something that would have survived being inundated with water about 9,000 years ago. And so one of the things they knew about that time period was that

caribou were the main source of food. And they also knew that prehistoric hunters made these really cool hunting structures, they're called drive lanes, and they would guide the caribou to these kill sites. And so John O'Shea, his collaborator, they thought, if they were made of stone back then, maybe we could find them underwater. So it's all these hypotheticals. But it helps to know where the caribou would go so that they can know where to look for sites. So the idea is if the caribou follow a certain path, then humans probably aren't far behind. And then it's the humans who are leaving these artifacts. Exactly.

And why couldn't researchers just find these artifacts the old-fashioned way? Yeah. So technically they did find the first one, the old-fashioned way, kind of. I mean, they use sidescan sonar. I think they had an underwater robot at the time, but there wasn't any AI. But regardless, the challenge was that Lake Huron is huge. And even though the land bridge offers this concentrated place, this corridor to look, it's still really long. It's about 90 miles long.

It's about nine miles wide. And then on top of that, you've got to go 100 feet underwater. And here's John O'Shea talking about, you know, this process. Underwater research is always like a needle in a haystack. So any clues you can get that help you narrow down and focus the kind of places you might look at, There's a real help to us. And, you know, John happened to know the premier, one of the premier people doing archaeological computer simulation.

His name's Bob Reynolds. He's based at Wayne State University. And so their idea was they'll create a computer model of the LAMBridge and then use AI to help predict sites. And how does this AI actually work? What kind of information were the researchers plugging into the model? Great question. Okay, so the first step is you've got to actually build the virtual Lambridge, the Alpina Amberly Ridge, and they use the actual topography. And then they start populating it with digital caribou. And that's the piece that has the artificial intelligence. So they create these caribou and they give them instructions, their computer algorithms. And the instructions basically tell them how to behave. A really simple one is caribou walk. Okay, so the caribou start walking. And then another simple one is be aware of obstacles and move around them.

Like, don't bump into rocks or each other. Another one is move in groups and break apart. And they just keep refining and refining it until the caribou start behaving more and more like real caribou. And what did it look like to watch the AI model in action? I'm picturing a sort of animated video of caribou walking around. But is that what the early models really looked like? They had some glitches. One of the researchers I talked to, she described the caribou looked like they were roller skating because they didn't exactly walk.

But, you know, they've kept developing it. And that's really a whole other story. Now they have an amazing virtual reality where they really look like caribou. But, you know, another glitch was Bob Reynolds, the computer scientist. at Wayne State that I mentioned. He talked about this other one glitch that was funny. Literally the first model, we let the herd run across the land bridge. And they did not have edge perception.

And so they kept dropping off the size of the bridge like Lemmings, you know. Oh, no. RIP to the AI Caribou, I guess. I know. I know. And so it's a perfect example of where, you know, they have to introduce a new algorithm and they basically give the caribou a new instruction, which says, hey, you got to present.

receive edges. So it's a lot of trial and error and refining. And how well is the AI working today? I mean, have the researchers actually found any real world evidence based on these computer generated paths? Yes, absolutely. They've found prehistoric hunting sites. They've found artifacts. Ashley Lemke, she's an anthropological archaeologist also on the team, and she's currently a professor at the University of Wisconsin, Milwaukee. And here she or she, she, here she is. We could ask, as archaeologists, like, how did you find a site? Or, like, how did you know where to dig?

And for me, I can be like, oh, well, like, artificial intelligence told me, you know? So how it works is the Caribou develop these optimal roots over time. They're going back and forth and back and forth. And there were a few spots that they went nearly every time. And they call these choke points. And so this was a really obvious place for archaeologists to go and look. And this is just one example of how AI helped. but this one particular choke point led them to the site.

They call it drop 45. It's the most complex hunting structure found in the Great Lakes to date. And there's a number of things there. There's a line of stones guiding caribou to a kill site. There was a fireplace with burnt earth, incredible 9,000 years old. And then there were also these really unusual small tools that were unprecedented for the region. And, you know, that's AI. AI helped them find that. And it's saved them time and money. And now that these amazing

little artifacts are being, little and large, I guess, artifacts are being found. What's next for the team? Yes. So keep looking. Yes, they've found some artifacts, but they've just scratched the surface. And with what they found, they've started to build an understanding about what the environment might have looked like and how people might have lived. But they've also found some totally new and, as I mentioned, unprecedented artifacts. Here's Ashley Lempke again. None of this matches the models we had about peoples in this region, which is really, it's really fascinating because then you have to go back and be like, all right, well, now we have

this new data. You know, what does that mean for what we thought about people that were living in the Great Lakes? You know, you kind of have to rewrite the story. So they keep looking. They keep researching and they keep researching. keep rewriting the story. And now that it's been proven that this sort of application for AI works, do you think it'll gain traction in the larger scientific community? You know, I don't know. I think it should, for sure. But Bob Reynolds, he's the main computer scientist. This is not the only project he's worked on.

So it's definitely something he's working on with other archaeologists, other scientists. But it requires a lot of strong collaboration between completely different fields. And I know that Bob specifically, he leans heavily on students at Wayne State to really help make the simulation and the virtual reality come to life. That's all the time we have for now. I'd like to thank my guest, Morgan Springer, editor of the Points North podcast at Interlocan Public Radio in Interlock in Michigan. Thank you for joining me. Thank you so much, Sophie. That's it for today. Lots of folks help make the show, including Ariel Zitch

Jordan Smudjick Diana Plasker and many more tomorrow we'll check in on the top stories from the week in science I'm SciFry producer Charles Bergquist thanks for listening

we'll see you soon