Pod Insights TED Radio Hour The future of our memories About this episode From creating "synthetic" memories to reviving ruined monuments, tech no longer simply stores the past—it can enhance it. This hour, we explore new ways to capture, share and even recreate our past. Guests include technologist Pau Aleikum Garcia, cartoonist Amy Kurzweil and digital archaeologist Chance Coughenour.
TED Radio Hour+ subscribers now get access to bonus episodes, with more ideas from TED speakers and a behind the scenes look with our producers. A Plus subscription also lets you listen to regular episodes (like this one!) without sponsors. Sign-up at: plus.npr.org/ted
See pcm.adswizz.com https://pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.
NPR Privacy Policy https://www.npr.org/about-npr/179878450/privacy-policy
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Manoush Zomorodi? This episode's sponsor is HubSpot, which offers the following message: What if instead of updating your CRM, your CRM updated you? HubSpot's self-updating CRM keeps your customer context current so you can win more deals. Learn more at hubspot.com. This is the TED Radio Hour. Each week, groundbreaking TED Talks.
0:24 Speaker 01 Our job now is to dream big.
0:25 Manoush Zomorodi? Delivered at TED conferences.
0:27 Speaker 02 To bring about the future we want to see.
0:29 Manoush Zomorodi? Around the world.
0:30 Speaker 02 To understand who we are.
0:32 Manoush Zomorodi? From those talks, we bring you speakers and ideas that will surprise you.
0:37 Speaker 03 You just don't know what you're gonna find.
0:39 Manoush Zomorodi? Challenge you.
0:40 Speaker 04 We truly have to ask ourselves, like, why is it noteworthy?
0:42 Manoush Zomorodi? And even change you.
0:43 Speaker 05 I literally feel like I'm a different person. Yes. Do you feel that way?
0:48 Manoush Zomorodi? Ideas worth spreading from TED and NPR.
0:55 Manoush Zomorodi? I'm Manoush Zomorodi. On the show today, the future of our memories.
1:02 Speaker 06 So it was 2014, and it was during one of the biggest refugee crises from the last decade here in Europe.
1:12 Manoush Zomorodi? This is Pau Aleikum Garcia. Pau lives in Barcelona, but a decade ago, he was in Greece helping relocate Syrian refugees who were escaping civil war.
1:26 Speaker 06 And I remember one conversation one night with this elderly woman. And she told me, Pau, I'm not afraid of being a refugee now as I have lived my life. What I'm afraid is that my grandkids will be refugees forever. And I said, how come? I don't think so. Like, they will find a new place, a new home. They will keep on. She told me, yes. The thing is that whenever they try to ask the question, where do I come from? They won't be able to answer it because we have lost our home. We have lost our neighborhoods, the streets where they were born. And of course, we have lost all our photo albums
2:11 Speaker 06 from our family, the diaries of my husband. Not only community and country heritage have been lost, but also our family, individual family history have been lost through this forced migration.
2:26 Speaker 06 And that kind of made us reflect on the role that memory, and visual memory, has in the buildup of our identity and how we construct ourselves. And that idea of reconstructing memories was stuck in us.
2:47 Manoush Zomorodi? Reconstructing memories, this got Pao thinking. Could he help people capture their memories in a photograph? Something to hold and see, even if their family albums, diaries, and letters were lost? Because Pao was a technologist, and so over the next decade, he and his colleagues at his design studio in Barcelona worked with AI tools to create something that they now call synthetic memories.
3:18 Speaker 06 A synthetic memory in the context of this project, of course, refers to a digitally reconstructed representation of a personal memory. And we use artificial intelligence models that, using words, prompting them with words, can generate certain images. But the goal is to find an image that can bring a face to stories of people that have lost these images, that have never been able to document part of their memories.
3:51 Manoush Zomorodi? Pao Aleikum Garcia explains from the TED stage.
3:56 Speaker 06 Memories are the architects of our identity. Memories remind us who we are. And visual memories are very important. We all have a picture of someone we love in our wallet, in our phones, in our homes. Visual memories shape our sense of self, and they can shape our sense of belonging to a specific place. They can teach us things from the past. And because all that, they can actually make us understand in a deeper level how we react to things. And like organic memories, which are formed and stored in the human brain, synthetic memories are visual memories from a person past, which have been never documented or
4:41 Speaker 06 lost, and that we can now, using GenAI, transform from a text or an oral description into an image.
4:50 Speaker 06 Is a new way of reconstructing a past that has been hidden to our eyes.
4:57 Manoush Zomorodi? Part of me wonders, well, why do we need a synthetic memory? Why do we need artificial intelligence to create a visualization of it?
5:06 Speaker 06 It is, I think, a matter of sharing memory with family members or members of their community. And what I have seen is that it kind of dignifies certain experiences in life. There is a lot of people that because the moments that they lived and some of these situations were very traumatic or very difficult, the only traces that they have of these experiences are psychological scars. There are no visible ways to explain what happened to them. And by reconstructing these memories, by documenting in this kind of subjective way
5:48 Speaker 06 what happened to them, you are also dignifying part of their own experience.
5:55 Manoush Zomorodi? From AI-generated memories to chatbots that let us speak to loved ones we've lost, technology isn't just storing our memories anymore, it's transforming how we interact with them. But what happens when the past doesn't stay in the past? Will it change our conceptions of ourselves in the here and now? Today on the show, the future of memory: how technology is changing the way we capture, protect, and even recreate our past. Back to Pau Aleikum Garcia.
6:35 Manoush Zomorodi? He says that synthetic memories can restore a fundamental part of what makes us human, as it did for a 90-year-old woman named Carmen.
6:46 Speaker 06 Carmen was one of the very early participants of synthetic memories. She came to the studio and we started, as we start most of the sessions, asking, what is your earliest memory? It was from 1941. She was six-year-old and her mother would pay another family so they could enter their house and go up to their balcony. What was particular about that balcony was that it was facing La Modelo Prison. During the time, during the Spanish dictatorship in Spain, it was a political prison. And her father, a doctor for the anti-fascist front, was a prisoner there.
7:28 Speaker 06 So the only way they could see each other was from that balcony to the window of the prison. And that was her earliest memory of him, between bars through that street. I asked her, Carmen, would you like to have an image of that memory? And she said, yes, of course. I would love to show to my family what I experienced, the things that I went through, so they can remember where we all come from.
8:01 Manoush Zomorodi? Just so that the listener can picture it in their minds, the photograph that you ended up creating with AI was of a girl in a dress, standing at a balcony. Her mother has, you know, this old-fashioned haircut.
8:16 Speaker 06 Exactly. She described very well how her mother was dressed, the haircut that she remembered. Also, the balcony, how was it? And kind of the general atmosphere of it. We went through several iterations. We generated tens of images. And then at some point, she saw two images. That was like, yes, that was it. That's how I remember it. And it was, I know, it was very visceral, really. It was something I didn't experience before, something like this. And it's a feeling that I have seen over and over afterwards with other people, when they see an image that was only in their mind
8:58 Speaker 06 and then it's externalized by reconstructing an image that other people can see now. And it's a kind of a release for them. It's like, oh, OK, I don't have to hold it in my mind only. Now I can show it to other people. And it's not only on me to keep that memory.
9:18 Manoush Zomorodi? But it's not very realistic looking. It has a painterly, sort of dreamy feel about it. It's in black and white. It almost feels like you can't quite grab it, much like a memory.
9:33 Speaker 06 Exactly. At the beginning, it was because the models that we were using were not good enough. They were very early models. But then, when we started to switch into more hyper-realistic models, we discovered that actually, the models that worked better, the models that created images that people felt more comfortable with, were the early models, models that were generated like this kind of blurriness. And it was because memories are also unstructured. And whenever you do a hyper-realistic image, people start to look at the things that were not exactly how they remember. And it's not so much about the factual accuracy of the image,
10:14 Speaker 06 it's more about the emotional embedding that you find in them.
10:19 Manoush Zomorodi? You've been doing this for over three years now. Who do you think this is most helpful for?
10:26 Speaker 06 So synthetic memories can be useful in several spaces. The first space is for communities that because forced migration, they have lost hard drive phones and photo albums. Another space that I think is relevant is in the space of dementia. So on the last two years, we have been actually starting pilots in different nursing homes and hospitals, transforming synthetic memories into a methodology within a bigger space of therapy, which is called reminiscence therapy. It's a therapy that uses old images and music from someone past to actually trigger these very visceral memories.
11:09 Speaker 06 And this has been proven to actually reduce anxiety and depression for a big number of people.
11:17 Manoush Zomorodi? You've been very upfront about your concerns, your ethical concerns around using this kind of technology or and helping people who maybe aren't, don't know how to use the technology do this. What are the worries that you are hoping to sidestep?
11:36 Speaker 06 The most obvious one is fake memory reconstruction, right? Like creating memories that never existed and then making someone believe that that happened. And this is, of course, is something that is not new. This has happened since very early software, like Photoshop, transforming images to show things that were not there in reality. I think now with this technology, what can happen is that can happen at scale. And that's why whenever we create synthetic memories, we always use models that you can very easily see that they are algorithmic generations, that they are not images taken
12:19 Speaker 06 with a photographic camera. The other thing is that we want these memories to be subjective and individual. We don't want them to be collective. I think there is kind of a danger on reconstructing collective memories as they can build community perspective that never happened. This is something that has been studied a lot. I think there are ways of using these technologies that can be good. But artificially intelligent is neither only a threat or just a simple opportunity. We have to learn to be OK with it being both at the same time.
12:56 Manoush Zomorodi? When we come back, more from Pau Alacum Garcia on the potential ethical problems with memory reconstruction. On the show today, the future of memory. I'm Manoush Zomorodi, and you're listening to the TED Radio Hour from NPR. Stay with us.
13:40 Manoush Zomorodi? Standards from Whole Foods Market. There are no antibiotics in their meat department ever. So carve up turkey and spiral cut ham with peace of mind. Shop organic sides like stuffing mix and cranberry sauce from 365 brand. And all their delectable pies are made with no high fructose corn syrup. Gourmet Thanksgiving flavors with the quality you want only at Whole Foods Market. This message comes from University of Buffalo, where scientists are working to protect the Great Lakes and freshwater supplies around the world from harmful contamination. Learn more about their research at buffalo.edu/NPR.
14:16 Speaker 01 This message comes from Google Workspace with Gemini. One subscription that does it all. Personalized AI in premium business versions of Gmail, Google Docs, Drive, and more. Try it free for 14 days at goo.gle/workspacewithGemini.
14:34 Manoush Zomorodi? It's the TED Radio Hour from NPR. I'm Manoush Zomorodi. On the show today, the future of our memories. We were just talking to technologist Pau Aleikum Garcia, who makes what he calls synthetic memories. Photographs created with AI to depict a moment from someone's past. But memories can be subjective. We know memory is fallible. It's unreliable. Have you seen, you know, is there concern that some memories, say from war zones or conflict areas, that people create them and they're not necessarily real, but they reinforce
15:17 Manoush Zomorodi? or perpetuate narratives that didn't actually happen, that maybe are not accurate?
15:24 Speaker 06 I think the danger is not on individual subjective experiences, but in general narratives. And I think that's one of the red lines for us. We always have to talk about individual subjective memory. It's not forensic truth. Actually, we had, we had these interviews with people that were tortured during the dictatorship in one of the police quarters here in Barcelona. At some point, they wanted to use these images. We we didn't recommend it. We say that it's not an image and it's not a forensic proof of anything. It's just another format of subjective representation of what happened.
16:09 Speaker 06 Because, you know, memories are undefined and are imperfect. Like, every time we remember something, we turn it different. Like, we change a bit, a part of it. And whenever we are creating an image with one of these models, we are not only reconstructing the memory of it, but we are also rebiasing it through the biases of the model that we are using. For example, here in Barcelona, at the beginning, we were using DALL-E 2 from OpenAI. And this is an inherently American algorithm, like image generator. So it has been trained mainly with American culture images. And whenever you try to generate images from Barcelona,
16:53 Speaker 06 from the 20s or the 30s or the 40s here, it's not very good at it. It's not very good at reconstructing how a home from a person in Barcelona was in the 30s. So we had to do, as a partnership with the municipal archive here in Barcelona, and fine-tune a model with a lot of images from the time in the city. So there are ways that you can somehow rebias and fine-tune these models to look closer to the cultural context of the person that is trying to recall a memory.
17:28 Manoush Zomorodi? You know, I'm thinking about the way my kids live right now. Everything is documented. They take photos of absolutely everything. And how much they think they remember a moment, they can go back and check whether they were accurate. How is the potential for using this changing as you get to younger and younger people who do have photos of everything?
17:48 Speaker 06 Yeah. I think this project is, right now, very focused on this window, that is closing of these elder generations that are starting to die. And with them...
18:00 Speaker 06 It's not only the body that dies, it's also all the experiences and memories and ways of looking at the world. And a big part of these memories and realities were never documented. And I think sadly still, for future generations, they will still go through experiences like massive migration or forced migration that will leave them without the capacity to hold into visual memories. And some of them, they will lose tragically like all these images. So I don't think this is something that only affects the elderly, but also in the future, can affect also a lot of the younger generation.
18:40 Manoush Zomorodi? That's Pau Aleikum Garcia. He leads the design studio Domestic Data Streamers in Barcelona. You can see his full talk at ted.com. On the show today, the future of memory. And now a story about three generations of the Kurzweil family and how technology has kept their family history alive.
19:05 Speaker 06 He was born in 1912 in Vienna.
19:11 Manoush Zomorodi? This is cartoonist Amy Kurzweil talking about her grandfather, Holocaust survivor Frederick Kurzweil.
19:18 Speaker 06 He was a very talented musician and he became a pianist and a conductor. And he had a really promising career in Vienna and was sort of, you know, looking forward to a life of being a professional musician in the music capital of the world.
19:33 Speaker 06 And in 1937, he conducted this choral concert.
19:41 Speaker 06 And there was an American woman in the audience who was a wealthy benefactor, who loved music, and she really loved his performance.
19:52 Speaker 06 And afterwards, she went up to him and she said, "I just am so impressed with you. And if you ever need anything, let me know. Here's my card."
20:02 Speaker 06 That was 1937.
20:06 Speaker 06 In 1938, the Nazis marched into Austin.
20:09 Speaker 02 Austria and took it over. And my grandfather realized very quickly, you know, sort of overnight, that he needed to get out of Vienna. And so he wrote to this woman and she ended up sponsoring him.
20:26 Speaker 02 And I grew up hearing this story and I think it took some time for me to realize just how significant and dramatic that moment of the concert and the decision to sponsor him was. Because he really didn't have any other way out. Music is really what saved my grandfather's life. And so he, yeah, he left in 1938. He left a month before Kristallnacht. And he settled eventually in New York. He met my grandmother and that's where history unfolded.
21:00 Manoush Zomorodi? And they had a son, your father. Tell people, if you don't mind, who your dad is.
21:06 Speaker 02 Yeah, my father is Ray Kurzweil, inventor and futurist.
21:10 Manoush Zomorodi? I've actually been in artificial intelligence since 1962. This is Ray Kurzweil being interviewed on NPR News in 2023. Which is actually a record. I don't think there's anybody else that's been in it for that long.
21:24 Speaker 02 He's, you know, known to the world as a kind of genius propagator of the singularity, somebody who has really fantastical ideas about the future of humanity and the rise of AI. But to me, he's just my dad. And he's like, you know, he's a really wonderful, sweet person who cares about his family.
21:45 Manoush Zomorodi? Ray Kurzweil and his wife had Amy after his father Fred had died.
21:51 Speaker 02 Yeah, so I, I never met my grandfather because he died in 1970.
21:56 Manoush Zomorodi? He died when I was 22, so.
21:58 Speaker 02 So really what I had from him were these, these stories and these fragments. Of a life, sort of filtered through my father. They said, oh, Fritz Kurzweil's in town. So she went to meet him. Six dates later, they were engaged. I wouldn't say my father talked about his father all the time, but I could tell that when he talked about him, there was great emotion and great longing and a sense that this is a person that he started to have a relationship with only at the end of his life, or they would talk and they would share memories. My father felt that was kind of wasted. He could have done more creative work and we talked a lot about that. And he was very enamored with that.
22:38 Speaker 02 I think what impacted me in my father's stories about his father was the absence and the longing and the sense of great admiration, that this was really someone who had dedicated his life to art and creativity and ideas. And that was really significant to my father and had a big impact on him.
23:00 Manoush Zomorodi? Amy's grandfather kept reams of artifacts and documents and letters from his life.
23:06 Speaker 02 Yeah, he was not able to bring very many things with him from Vienna when he immigrated. But documentation is very important in this German Austrian tradition. And so my grandfather came to America with this sense of the importance of documentation and the importance of information to save your life. So in America, he saved copies of all the letters he wrote. He saved copies of his programs when he would do musical performances. He just saved and saved. And so my father inherited this trove of salvaged documents and artifacts from my grandfather.
23:51 Speaker 02 My father's life.
23:53 Manoush Zomorodi? I have actually many boxes of his projects. And we have everything he's ever written. Ray came up with an idea to do something radical with all these documents. And he drafted Amy to help him.
24:08 Speaker 02 In 2018, when large language models were starting to gain capability, my father had this idea, which is that if he could collect all the writing from his father, he could create a chatbot that writes in my grandfather's voice.
24:27 Manoush Zomorodi? A bot that would respond via text as though it was her grandfather.
24:33 Speaker 02 And so what this required was collecting all the original writing from my grandfather, which was housed in the storage unit, and marrying it with natural language technology. And that would create this sort of first creation that represents my grandfather's identity in text. And so my role in the project was to collect these documents and transcribe them. And eventually that built this large language model, this Fredbot, as we called it. Back in 2018, this seemed very sci fi.
25:07 Manoush Zomorodi? Amy Kurzweil picks up the story from the TED stage.
25:10 Speaker 02 But I wanted to talk to my grandfather because he, like me, was an artist. I wondered, could I get to know him? Could I even come to love him, even though our lifespans didn't overlap?
25:27 Speaker 02 This was a selective chatbot, meaning it responded to questions with answers from the pool of sentences that Fred actually wrote at some point in his life. The more examples of Fred's writing we could find, the more dynamic the experience of chatting the bot would feel.
25:43 Speaker 02 Finally, I sat down to chat with this new intelligence, an algorithm commanding over 600 typed pages of letters, lectures, notes, essays, and other written documents from the grandfather I never met. When I asked about Fred's dreams, he told me about the challenge of keeping his new orchestra afloat. When I asked about Fred's anxieties, I learned about the stress of being a new father while working so hard. When I asked about the meaning of life, Fred wrote about the joy of working with other musicians in pursuit of beauty, and he wrote about the highest aims of art.
26:25 Manoush Zomorodi? You ended up writing and illustrating a book called Artificial: A Love Story about you and your father's quest to take your grandfather's archives and digitally turn them into what you called Fredbot. But just so people can picture it, there is a page in your book depicting you and Fredbot texting each other. And it was less of a conversation, more like you would ask a question and he would respond with a line pulled from the archives that was relevant. It was almost like you were reciting poetry to each other.
27:01 Speaker 02 Yeah, that's a nice way of putting it. The first time I chatted with Fredbot was, I believe, in 2018. And I had to actually go to a physical place to chat with him because it was using technology that was not publicly available. Of course, today, I can create many different Fredbots on all these different platforms. I can create one using NotebookLM. I can create one using ChatGPT. And there's different techniques for prompting these large language models to get them to engage with me as Fredbot if I give them the corpus of text that we built.
27:38 Manoush Zomorodi? Oh, right.
27:39 Speaker 02 But back in 2018, this stuff was not as productized. I would ask a question and it would offer something that was related. And then I was faced with this gap between what I asked and what it offered. And then I had all this space to kind of imagine where that line came from. And I could really enter into my grandfather's experience because of that gap. And the metaphor that I use in my TED Talk is this idea of time travel, that when I would ask a question that was relevant to me in my own life right now, like, tell me about Hannah, my grandmother. The Fred Bot would come back with a passage that just took me to this moment
28:23 Speaker 02 when my grandmother was in a car accident and he was worried about her. And then suddenly, out of nowhere, I would just be entered into this past situation where my grandfather was, you know, writing to a colleague about how he was, you know, having to take care of my grandmother and she was lying in bed for weeks. And I think just the impact of suddenly being plunged into a totally different moment in time was so, so much more interesting than if the chatbot had tried to pretend that he was around right now.
29:00 Speaker 02 I had wondered if this project would feel like a resurrection. But rather than bringing my grandfather from the past into the present, it felt like I was the one time traveling, visiting him for a moment at different points in his life. It felt like the kind of imaginative travel I do when I'm cartooning. When I'm cartooning, I'm always thinking about how I could possibly represent a person fully. And the answer is, I can't. Similarly, I know how many aspects of my grandfather can't be captured by digital text alone. There's his body, how it moved and how it felt. There's his music and all the ineffable aspects of his performance.
29:45 Speaker 02 And of course, there's everything he thought but didn't write down. What would we have to do to be able to capture all of this?
29:57 Manoush Zomorodi? To be clear, when you talked with Fred Bot, it would respond in written sentences, not his actual voice.
30:04 Speaker 02 Written sentences, yeah. There was no voice. And something I should say about the voice is that, of course, voice recreation is now possible, but we have no recordings of my grandfather's voice. So we can't reproduce his voice. We can't have, you know, an algorithm say anything new in his voice.
30:24 Manoush Zomorodi? Does that make you sad? Or are you kind of like, it's kind of beautiful that we just can be immersed in his inner thoughts from his writings?
30:32 Speaker 02 I think in the case of my grandfather, what's most significant is recordings of his music, because that's what was important to him.
30:42 Speaker 02 We do have recordings of his music. We have him playing piano. We have him conducting. Curzwell is coming onstage to lead the New York College of Music Chorus. For me, I think the piano is the most important because I see that as his greatest passion and the place where his personality really shined. I wish that I could see it, but just like on the idea of having some things that are gone forever, I think that's sad, but it's also part of what makes this whole experience precious, that there is something about absence that that kind of calls us to attention.
31:23 Manoush Zomorodi? I want to go back to the motivation for your dad and your understanding of it. There's a page where your father talks about the idea of the Mexican concept of death, that there are two of them. Can you tell us the story? He told you, I mean, it's similar to the movie Coco, right?
31:43 Speaker 02 Yeah, my favorite Pixar movie, Coco, which explores this idea of the two deaths. The first death is the death of the body. The second death is the death of your legacy. And what that means is when people forget you, when they stop talking about you. And that was a framework that I was putting on this project of my father's, that there was a way in which he was keeping his father alive through allowing us to engage with his memory. And, you know, I was, in my book, I was grappling with these questions of like, what is a person? Where does a person begin and end? Is there a way in which we think about the Fred Bot as an extension of my grandfather's identity, a part of his identity?
32:28 Speaker 02 And I came to see the Fred Bot as an artistic representation of my grandfather.
32:35 Manoush Zomorodi? In a minute, more with Amy Curzwell on the limitations and beauty of using technology to capture who we really are. Can we be reduced to ones and zeros? On the show today, the future of memory. I'm Manoush Zamirodi, and you're listening to the TED Radio Hour from NPR. We'll be right back.
33:07 Speaker 01 Stay tuned for the following announcements and messages from our sponsors.
33:11 Speaker 05 This message comes from Whole Foods Market. Host Thanksgiving with industry-leading quality standards from Whole Foods Market. There are no antibiotics in their meat department, ever, so carve up turkey and spiral-cut ham with peace of mind. Shop organic sides like stuffing mix and cranberry sauce from 365 Brand. And all their delectable pies are made with no high-fructose corn syrup. Gourmet Thanksgiving flavors with the quality you want, only at Whole Foods Market. This message comes from University at Buffalo, where scientists are working to protect the Great Lakes and freshwater supplies around the world from harmful contamination. Learn more about their research at buffalo.edu/npr. This message comes from Schwab.
33:54 Speaker 05 At Schwab, you can get everything from self-directed investing to full-service wealth management, all in one place. No matter your investing goal, life stage, amount to invest, or know-how, you can invest your way with Schwab.
34:10 Manoush Zomorodi? It's the TED Radio Hour from NPR. I'm Manoush Zomorodi. On the show today, the future of memory. And we were talking to artist Amy Kurzweil, whose father is Ray Kurzweil, the futurist famous for his ideas about the singularity, the merging of man with machine, so that we could, potentially, live forever. Amy thinks her dad's ideas can be interpreted in many ways.
34:40 Speaker 02 The way that I interpret my father's immortality aims is actually more about the present and about always trying to keep living. And living doesn't just mean not dying. Living means enjoying your relationships. It means seeking meaning and really trying to push your life forward towards some creative aim. I find that to be inspiring.
35:08 Manoush Zomorodi? I love this quote so much. You write, "A person is ever-changing, but surely something makes you the same person over time." My father says, "A person is a series of patterns." And then there's a really funny section where your mom plays different voicemails from your dad that he had left like over weeks. And every time he leaves a voicemail, he starts the message the same way. "It's me." "It's me."
35:37 Speaker 02 Yes. Yeah, I love that you brought this up. My mom has a great eye for humor. And she, she saved all these voice messages from my father because he was doing a lot of traveling and he would always call and cough. "Hi, it's me." And we just sat and laughed. And I think people can relate to that because when you're close to somebody, you have this really felt sense of who they are and the things that they do. And it's just funny. And yeah, I think patterns are what connect our identities over time. So I mean, my father says, "A person is information." I think when people hear that, they hear, "A person is reducible to ones and zeros."
36:17 Speaker 02 But actually, like in philosophical literature, the idea that we are patterns of information is, is, you know, pretty common. Lots of people think that. So the idea that we are the shapes of our life, the sort of stories and preoccupations and relationships that we are living out. And the patterns of our lives could be incredibly, transcendentally beautiful and complex. But nonetheless, they are patterns. They are things that we can observe.
36:50 Manoush Zomorodi? This episode is all about how we will be thinking about the past. And you say in your talk, "AI has a special role to play in the mission of memory."
37:04 Speaker 02 Yeah. I mean, I think what AI can help us do is navigate incredible amounts of information. Something that I discovered in my writing my book is that the amount of information that a person is, is so vast as to be potentially infinite. And of course, the process of memory is the process of choosing and selecting what moments and what details from a human life are important. And so because AI has this capacity to process much more information than we consciously can, and because AI doesn't have emotion, it just has a different method for selecting
37:47 Speaker 02 what's significant in its series of patterns that it's recording. And I think it's really promising for sort of helping us keep archives alive and keep archives navigable. And I think when we have a sense of where we've come from, our own identities feel more solid.
38:09 Speaker 02 I did not come to see the chatbot of my grandfather as replacing my grandfather. I came to see it as one way to interact with his legacy. As somebody who has spent their whole life trying to document people, I can assure you that people are much bigger and weirder than any one depiction or any one moment in time can possibly evoke. AI swirls our conception of time and space. It can remix and extend our identities. These technologies are animated portraits. They are one part of our true immortal selves. Seen this way, AI, like cartooning and all good artistic endeavors, could help us appreciate the vastness of humanity if we let it.
38:55 Speaker 02 Thank you.
38:58 Manoush Zomorodi? That was Amy Kurzweil. She's a cartoonist at The New Yorker Magazine and the author of two graphic memoirs, Flying Couch and Artificial: A Love Story. You can see her full talk at ted.com. On the show today, the future of memory. How technology is changing how we remember our past. So far, we have focused on preserving personal memories. But what about collective memories of a people and their culture?
39:30 Manoush Zomorodi? Take St. Sophia's Cathedral in Kyiv.
39:38 Speaker 06 The heart of Ukraine, the main church in Ukraine, more than 1,000 years.
39:44 Manoush Zomorodi? This is Yurko Prepodovny. He says for 1,000 years, Saint Sophia, with its stone towers and 13 gold cupolas, has stood at the center of historic Kyiv.
39:58 Speaker 06 It's not simple structure. It's from the underground few rooms to the main hall. Corridors, tunnels with arches inside and a lot arches, a lot cupolas inside.
40:12 Manoush Zomorodi? The cathedral is a rare surviving example of Byzantine architecture and a UNESCO World Heritage Site. It's also considered the heart of Ukraine for the art within its sanctuaries that depict the earliest moments of Ukrainian history.
40:28 Speaker 06 Saint Sophia has the oldest mosaics fresst on the walls. That walls include all our history like Ukrainian. It's the layers and layers of the centuries.
40:50 Manoush Zomorodi? The capital of Ukraine under a new wave of Russian drone attack.
40:54 Speaker 02 A barrage of over 150 Russians.
40:56 Speaker 03 Military officials say Russia targeted the city with drones.
40:59 Speaker 06 Every building in Ukraine in danger.
41:04 Speaker 06 Saint Sophia had missile attack 600 meters from them.
41:10 Manoush Zomorodi? Saint Sophia and hundreds of other historic buildings have been damaged or destroyed in Ukraine over the last few years, which is why Yurko and his team went to visit Saint Sophia and scan it.
41:25 Speaker 06 Yeah, we scanned everything. Every morning.
41:29 Manoush Zomorodi? We woke up and we scanned that building for one week. Yurko is a co-founder of Skyron, a Ukrainian organization that does digital preservation by creating 3D digital versions of heritage sites. He's already scanned the city center of his hometown of Lviv, and he and his team are now scanning buildings across the entire country. Now, we scan more than 50 per year. In the past, when war, natural disasters, or the passage of time damaged or destroyed buildings, all that were left were sketches, photographs, and memories. But now, high-tech 3D scans like the ones Yurko is doing preserve every physical aspect
42:14 Manoush Zomorodi? of these places in detailed digital files.
42:18 Speaker 01 What I've witnessed firsthand is the way that technology can safeguard the memories and heritage that we have of our shared humanity.
42:26 Manoush Zomorodi? This is Chance Coconhour.
42:28 Speaker 01 I'm a senior program manager at Google Arts and Culture.
42:32 Manoush Zomorodi? Google is one of Skyron's partners in Ukraine, offering both tech and financial support. And Chance is an expert in scanning technology. He started out as a Mayan archaeologist and first learned about this technology in the 2000s. But it wasn't until 2015 that he was convinced of its potential in Iraq.
42:54 Speaker 01 So in 2015, ISIL, or the Islamic State of the Levant, they entered the Mosul Museum in Iraq. This morning, ISIL claims to have destroyed priceless pieces of history. A video posted yesterday. They recorded themselves destroying the museum collection.
43:10 Manoush Zomorodi? The Mosul Museum reportedly housed more than 170 genuine antiquities.
43:14 Speaker 01 That video went viral. I remember when it was shared online.
43:19 Manoush Zomorodi? People around the world were shocked at the loss of these ancient artifacts. And it made Chance wonder if he could, at the very least, go back and create a digital record of them for posterity.
43:31 Speaker 01 If there were a number of photographs of these artifacts...
43:35 Speaker 05 Maybe we could create a digital copy or 3D model that could bring that artifact back from being lost, at least in a digital form, to preserve its memory for future generations. It was like an aha moment that we thought, why not at least try? And at that point, to be clear, we had no idea that anyone had any photographs of these objects. So really, it was just an hypothesis.
44:03 Speaker 06 And so, two weeks after we saw this video, we started the project called Project Mosul.
44:10 Manoush Zomorodi? Jan Kokenauer explained from the TED stage in 2016.
44:14 Speaker 06 The key to this technology is called photogrammetry, and it allows us to use two-dimensional images taken of the same object from different angles to create a 3D model. I know you may be thinking, this sounds like magic, but it's not. What the computer can do is it can detect similar features between the photographs, similar features of the object. Then, by using multiple photos, it can begin to reconstruct the object in 3D.
44:45 Speaker 05 It's very similar to how our eyes actually distinguish depth when we look at objects that are near or far from us. And so, it's really the application of that. It's comparing one photo with another, and then another, and then another in sequences. And then, algorithms piece those together to form a three-dimensional representation of that object.
45:06 Manoush Zomorodi? So, how do you go about tracking down photographs? I mean, tourists who'd visited Mosul, is that where you start?
45:12 Speaker 05 Yeah, you know, we're talking about a location that had been closed to the public for a number of years. You know, it was a zone of human conflict for decades, actually. And so, we just kind of placed an open call to find anyone out there. That may have these photographs. And a US servicewoman had happened to have visited the museum and taken photographs years prior. And I would say the most famous of the artifacts, she had taken 12 photographs of the Lion of Mosul.
45:46 Manoush Zomorodi? The Lion of Mosul was an eight-foot-tall statue carved around 3,000 years ago that stood guard at an Assyrian palace.
45:56 Speaker 05 They were perfectly taken from the different angles. So it was quite an incredible, we were just like, well, let's put the 12 photographs when we got them in. We were like, let's just test it. And we were like, wow.
46:10 Manoush Zomorodi? After Chance got the photos, it came back to life on his computer. He could even 3D print it, a perfect reproduction, small enough to sit on his desk.
46:22 Speaker 05 And at the time, we were testing software that create 3D models through photogrammetry. And so we tested it with many of them. And we created the 3D model in only a couple of minutes, even to the point where you can zoom in to see the original text that was inscribed on the side of this ancient carved lion.
46:47 Manoush Zomorodi? An amazing thing lost to history. You preserve it in some ways digitally, you can print out a 3D version of it. But that doesn't mean that it remains in people's minds, right? Like, how do you make it possible for people to even know that it ever existed?
47:06 Speaker 05 Yeah. So in 2019, the Mosul Museum hosted its first exhibition after the occupation had ended. What they wanted to do was to have a celebration of art and music and bring the community together. And we were invited to participate. So we sent a number of these 3D prints to the museum. They were actually placed alongside artworks by local artists that had painted during the occupation. So this was kind of an example of connecting the past to the present. But we also used these 3D models to place them inside of a virtual museum. You could go on a VR tour. And so many people became
47:49 Speaker 01 Aware of incredible heritage from this region. And the object itself, as a digital twin, is now accessible to zoom in, rotate and see in a way that would not have been possible even if you visited a museum in Mosul.
48:10 Manoush Zomorodi? It's been a little while since you gave your talk and since then you've joined Google. You've continued doing this work and the technology has advanced a lot.
48:21 Speaker 01 Yes, it really has. The technology has expanded and developed so rapidly. The technology that makes photogrammetry possible is the same that's used for AI today. It's now accessible to so many more people that it's much cheaper to do. Photogrammetry is being done on small mobile devices. A drone can fly around outside of buildings or even inside of buildings. But essentially, we have vastly expanded the number of 3D objects and monuments that we're putting online, along with our partners that are doing this globally. With our partners in Ukraine, for example, we've launched more than 200 more 3D models
49:03 Speaker 01 that range not only from landmarks, but in addition to that, museums that we've worked with could essentially take artworks and hang them in a virtual gallery. People around the world can virtually walk around inside of a museum.
49:18 Manoush Zomorodi? With accessing these things virtually or collecting them online, is there a part of you that feels, I mean, obviously that that's a wonderful thing to preserve it in some respect, but we are missing out on the tactile experience of seeing art
49:38 Manoush Zomorodi? and architecture right in front of us, which is an extraordinary feeling?
49:44 Speaker 01 I fully agree. I don't think it will ever replace the tactile way of being able to experience art in person. For me, the reason I became a Mayan archaeologist is because I walked into
49:58 Manoush Zomorodi? To an ancient Maya city, had some interest in history at one point in my life, and I was just blown away at the experience of being there. Mm hmm. But I'm still absolutely worried that things may be lost to natural disasters, climate change, time, human conflict. And what I think digital technology offers us is a means to digitally protect and document the world's art and culture in a way that it's accessible to people, for people to know about it, and hopefully, they have the opportunity to visit and see it themselves.
50:37 Manoush Zomorodi? That was Chans Kokonhour. He is a senior program manager at Google Arts and Culture. You can see his full talk at ted.com. Many thanks also to Yurko Prabodobny. His organization is called Skyron. To see their scans of St. Sophia and other historic sites, you can go to goo.gle/ukraine. And before you go, a big thank you if you're one of our new NPR Plus supporters. We so appreciate you signing up. Your support helps fund our work. And if you're thinking about joining, it is really easy. Just go to plus.npr.org.
51:21 Manoush Zomorodi? This episode was produced by Rachel Faulkner White, Katie Monteleone, and Matthew Cloutier. It was edited by Sanaz Meshkhanpour and me. Our production staff at NPR also includes James Delahoussaye, Hersha Nahada, and Fiona Giren. Our executive producer is Irene Noguchi. Our audio engineers were Robert Rodriguez and Gilli Moon. Our theme music was written by Ramtin Arablouei. Our partners at TED are Chris Anderson, Roxane Hylash, Alejandra Salazar, and Daniela Bellarezzo. I'm Manoush Zamirodi, and you've been listening to the TED Radio Hour from NPR.
52:00 Speaker 05 Stay tuned for the following announcements and messages from our sponsors.
Previous Your brain is the next tech frontier Next The shocking power of tiny things Page updated November 5, 2026 · transcript and summary are machine-generated
