Tony Marx on the Future of Libraries, AI, and Trust

30 Jul 2025 · 52 min

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Podcast Episode Summary: Tony Marx on the Future of Libraries, AI, and Trust

Podcast Overview Title: Possible Hosts: Reid Hoffman and Aria Finger Description: The podcast explores a vision of the future where technology and collective effort lead to positive societal change, focusing on discussions with leaders across various fields about the transformative power of technology, especially AI.

Episode Specifics Episode Title: Tony Marx on the Future of Libraries, AI, and Trust Description: This episode features a discussion with Tony Marx, president and CEO of the New York Public Library, about the evolving role of libraries in the age of AI, the importance of trust in information sharing, and how civic institutions can embrace technology while maintaining core values.

Key Themes and Discussions

  1. Introduction and Background
  2. Hellos and Intros (2:54): Tony Marx shares his journey from South Africa to leadership at Amherst College and ultimately to the New York Public Library (NYPL).
  3. Educational Impact (13:30): Marx discusses the transformative role of a year of quality education in overcoming systemic barriers.
  1. Evolving Role of Libraries
  2. Elimination of Fines (8:43): Marx highlights the removal of library fines to encourage access for all, particularly low-income families.
  3. Digital Transformation (16:20): Discussion on how libraries adapt to digitization and technology.
  4. Community Function (33:54): Emphasis on libraries as essential community hubs for learning and connection.
  1. AI's Role in Libraries
  2. AI Integration (17:42): Exploration of how AI tools can enhance library services and access to information.
  3. Navigating Hallucinations (23:08): Marx addresses concerns about misinformation and the reliability of AI-generated content.
  4. AI as Epistemology (27:57): Discussion on how AI may shape societal understanding and knowledge.
  1. Future of Work and Education
  2. Human Costs of AI (39:41): Marx discusses potential job displacement as AI transforms industries.
  3. Educational Gaps (42:43): Exploration of whether AI can help close educational disparities and provide access to personalized learning.
  1. Trust, Reliability, and Community
  2. Trust in Libraries (42:43): Libraries as trusted institutions and their role in navigating misinformation in the digital age.
  3. Role of AI and Libraries (39:41): The importance of libraries in maintaining trust and providing accurate information amid the rise of AI.

Key Takeaways

  • Libraries as Civic Institutions: Libraries are positioned as crucial public spaces for equitable access to information and community engagement.
  • AI's Dual Role: While AI offers tremendous potential for enhancing services, it also poses challenges in terms of misinformation and reliability.
  • Trust and Truth: The discussion emphasizes the importance of trust in institutions and the potential of libraries to lead in this area, especially in the face of technological advancements.

Conclusion Tony Marx's insights reflect a hopeful vision for the future of libraries as they integrate AI while preserving their foundational values of trust and community. The episode invites listeners to consider how technology can be harnessed to create equitable access and foster a society where information is shared responsibly.

For more information on the podcast and access to transcripts, visit [Possible Podcast](https://www.possible.fm/podcast/).

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:03Libraries are a great tool for sharing information that is trustworthy. They're the most trusted institution in this town or pretty much any town. There's no regulation because nobody ever thought our library would change. That means we can innovate at scale with people who trust us. Because the world we live in is getting less and less believable. So if there's a way we can use AI and its related tools to do more of what we're in the business to do and do it truthfully, that's our biggest asset, perhaps, more than the books.

0:54Hi, I'm Reid Hoffman. And I'm Aria Finger. We want to know how together we can use technology like AI to help us shape the best possible future. We ask technologists, ambitious builders, and deep thinkers to help us sketch out the brightest version of the future. And we learn what it'll take to get there. This is possible.

1:20When people think about the future of knowledge, they might picture cutting-edge AI models. digitized archives, or massive datasets. But behind those innovations are a few age-old questions. Who exactly gets to access knowledge and how? And who decides what gets included? For centuries, libraries have been a big part of the answer. But as books become e-books and search engines act as our librarians, the role of public libraries is shifting fast. At the same time, the need for trustworthy, equitable access to information has never felt more urgent. Here with us today to talk about that evolution is Anthony Tony Marks, president of the New York Public Library.

2:05Before leading one of the largest public library systems in the world, he was president of Amherst College, where he worked to diversify higher education. Under Tony's leadership, the New York Public Library has done everything from lending Wi-Fi hotspots to joining national AI initiatives. His team has helped transform the library from a traditional place of study into an innovative civic platform for lifelong learning and community. We talk with Tony about how technology is reshaping public institutions and what it means to not only preserve knowledge, but make it accessible to everyone. Here's our conversation with Tony Marks.

2:53So, Tony, thank you for hosting us for this. The New York Public Library is an awesome historic institution. You have been a great leader for a long time. Doing this in this amazing room is an extra. So welcome to Possible, and thank you for joining us. That's great to be here, Reid. Thank you. You know, you're a lifetime New Yorker. You know, grew up, went to public school here, have had a number of your different jobs here. But, you know, from South Africa to your studies on Latin America, tell us a little bit about your biographical journey. So I grew up, as you said, grew up in New York. My dad didn't go to college.

3:27I was lucky with the great public schools I went to and then with amazing colleges that got me to think. And, you know, I was always aware that my parents as kids had fled the Nazis in Germany. So I think that was playing in my head. When I got to college, South Africa was the big topic on campus. I knew nothing about it. It never did, had no connection to it. But I got taken in by that moment. And then after working in Philadelphia for a couple of years after college, I said, it's time to have an adventure. Friend of the Ford Foundation put me in touch with the best education person, anti-apartheid education in South Africa.

4:04And I found myself living there for most of, you know, a total of about three years in the middle of what would become a civil war in the 1980s. People I was working with were dying for what they believed in, and I'd never seen that. And the place I went to work for, the education group, said, we'd like to start a college that would send a few hundred students at a time onto the white universities who said, oh, we'd love to, but we can't find any qualified because apartheid had made sure of that. So we created a one-year residential college for 200 students a year in Cape Town and Johannesburg.

4:44And they went on to gain admissions and do very well at the great universities of South Africa. The lesson that changed my life was one year of quality education can undo 12 years, whatever, of apartheid, purposefully terrible education. that the human mind is capable of recovery and soaring. I came back, I wrote a dissertation about South African politics, the fights within Black politics, and then was teaching at Columbia. And then, sort of amazingly, someone put my name in for the presidency of Amherst. Tony, can you tell us how you made the decision to leave Amherst and take this job at the New York Public Library?

5:31I was, you know, happily doing that. I did that for eight years. I did miss New York, I confess. And then the library came a calling. And I, again, had little idea of what I was getting into. I've been very lucky. I've had these, like, two amazing jobs where I can sort of feel like I'm doing good in the world through education and learning. And, you know, the library, so very quickly, I mean, it's the largest library system in the country. It's the most used public research library on the planet, mostly because most of those are national libraries or university libraries. Almost all of them are.

6:09Libraries are amazing. They're the most trusted institution in this town or pretty much any town still. There's no regulation because nobody ever thought a library would change. Why worry about it? That means we can innovate at scale with people who trust us. That, like, never happens, right? That's designed never to happen. And very quickly, we've said, okay, the branches are in every neighborhood, including the poorest. Let's use the trust and the audience and say, okay, in addition to books and amazing librarians and heat and air conditioning, which is important for people, So let's turn these into education centers in every neighborhood, from pre-K to after school to homework help to college guidance to English language for immigrants, computer skills for everyone.

7:02And we moved from almost none of those to 2 million spots a year at this point to create opportunity through learning. When I got here, the place was still trying to figure out, I'll put it nicely, what to do about the digital world and whether we could hide from it. And so we said, okay, let's aspire rather than hide. And the aspiration now is any book you want, you should be able to get online if you prefer it that way. And if you're looking for the research library, any book you want should be available to you any place on the planet for free through your library. That means a lot of issues around intellectual property, a lot of digital issues.

7:48I have amazing colleagues. You know, that always makes all the difference for me, at least. Well, Tony, I'm especially excited to be here. My family moved to the Bronx when I was four. So I got my first. Where in the Bronx? I lived on the campus of Riverdale Country School. I was a faculty brat. I thought everyone had free tennis courts and a swimming pool. My kids grew up on the Amherst campus. You could compare notes with them. Absolutely. No, it was it was idyllic. And I got my first library card at age four from the New York Public Library. The first time I was allowed to walk by myself, I was eight and we had overdue books.

8:22And my mom was like, Aria, do you think you know where the library is? Can you take the books? Before the fines accumulate. I put the books in my backpack and walk to the New York Public Library and return them. And then the next day, of course, I was like, Mom, can I go get pizza? And she's like, you can't walk by yourself. I was like, no, I can. Remember when I needed to return the library books? Did you get to go for pizza? I did. I did. So it was a formative experience. So I will tell you, in case you don't know, I was shocked when I got here that we were gathering revenue from fines. And the problem of that, of course, is people think it's a necessary discipline to get the books back.

8:56Not true. The people bring the books back regardless. We know that now. What it does do is it keeps poor folks from coming to the library. Literally, I had people say to me, my parents wouldn't let me come to the library at all. because they were afraid I would borrow books and not return them and they couldn't afford the fines. We eliminated fines. Which is amazing. We're not in the book fine business. We're in the encouraged to read business, right? So I was actually listening to a podcast of yours, having you explain when you went to the Inwood Library, your childhood library, when you were sort of deciding to take this job, literally weeping on the subway this morning, coming in to prepare for this interview because it was such a beautiful time.

9:39And so I would love for you to talk about going to the Inwood Library and what you saw there that made you think, I have to take this job. So the library reached out. I'd really never thought of it as a place to work. So the library came a-calling. I was ready to come back to New York. So I went to visit the library I grew up in. And I wandered around and I saw, you know, people running businesses out of it, you know, kids studying in it, a teacher looking at books, tragically three levels below grade to what to assign to their class. And eventually I found myself in front of a guy from Latin America or Central America.

10:16I think my Spanish isn't great, but it's good enough. And I asked him what he was doing. And he said, well, he lives in an apartment with like a two room apartment with like eight or 12 people and there's no quiet and there are no books. And he comes to the library every day at lunchtime from his work to take a break. And I said, that's fabulous. What are you reading? And he sort of looked sheepishly at me. Well, I can't read. But he said, come over here. And he took me to a table where there were these big picture books I remember from my childhood of like Renaissance paintings. And he said, I come here every day to look at angels.

10:58And he pointed to an angel. And I was like, okay, I need to do this job. Some years later, it became obvious that that physical library was past its prime. And people said, you know, if we built a new one, we'd get a better library, a bigger library. You know, we could do a temporary library in that neighborhood's case in the meantime. And I said, great, you know, I've been wondering. why we never have used the air above our libraries. So they all, everybody, you know, who knows what they're doing, did a lot of work, came back and said, well, actually we could. So we built a brand new library and a senior center, I think it is, and a K-12 center, and above it, 175 units of 100 % affordable housing.

11:41Never happened before. Just never sort of occurred. So now my old library is this amazing new library. And we are now working every time we can to follow this model because we have the air. Why not monetize it for the library? And if that helps New York City with one of its most other pressing problems, God bless. I mean, I literally almost didn't move because I was moving 10 blocks further away from my local library. So the idea that you could live in a building and send your kids downstairs to the library is so beautiful. It's amazing. But it's also, that's because libraries are collections of books and computers from people who don't have it.

12:27I hope we'll talk about that as well. It's also community. We are human beings. We hunger for community. And the world we are building step by step is taking it away from us. And libraries are our third place, are our community. And they're going to become, I mean, they're still essential for all the assurance of books and education programs and digital and all that, but just the sense of sitting with each other in a world in which we are no longer capable of sitting with each other. Well, when I was growing up, I so much valued libraries. I would literally walk an hour to get to the nearest library for me in Berkeley was an hour walk.

13:12I know Berkeley. I was in the hills and I wasn't in your downtown and so on. So to get to the nearest public library branch, which was the North Berkeley Public Library branch, and then I would stay all day there because it was like I'm in heaven with a whole range of books. My mother used to park me at the public libraries because she didn't want me to come shopping. So let's start getting into the modern world. We'll obviously end up at AI, but let's start with kind of, you've got, I think it's 54 million books in the collection. I think 57, but okay. It depends on the date, right? 57, but who's counting?

13:45How do you make decisions about what to digitize? Basically, we collect and preserve anything we can within reason in the research library. We're not the Library of Congress. We don't get it for free. We have to make some choices. My colleagues have a pretty good idea of what the demand is like to be. And of course, we go way beyond it. We have the long tail. And there we will digitize based on two things. risk, meaning that we're going to lose it because we have old collections and things on tape, and demand, right? In the circulating collection, we can buy books as e-books, as well as buy physical copies and audio copies now, also growing market.

14:34And we try to keep all those based on demand, knowing that they will cycle out as something moves from the circulating library to the research library, one or two copies only, because we only have so much space. That's another constraint. But it seems to work, and demand is great. I mean, we used to say, this is what we think everyone should read. We don't do that anymore. We say, who are our patrons and our users, and what do we think they will want, right? You know, we are the library for the most diverse city on the planet. Our collections are also diverse. It looks different in that neighborhood.

15:10It moves around based on requests. That part we're really good at. I've never heard anyone say, oh, you missed one. Any surprising, like, insights from the demand side? A lot of romance novels, a lot of manga for, you know, young teenagers. Whatever you want. That's our job. Not to be judgmental, but to make it, to offer it, to make it accessible. By the way, one of the funny things I learned early GPT, I think it was GPT-3, they actually factored out a third of the romance novels because in the corpus was so heavy in romance novels. Oh, I thought you were going to say so repetitive in romance novels.

15:50No, just to try to balance it out was like the only editing that was being done was let's randomly take out a third of the romance novels. And I think they're not doing that anymore, but it was kind of a funny thing. Well, we'll come back to this. Remember, our job is not to feed AI. Our job is to feed the minds of human beings, right? This isn't the only use of libraries also, but like I think of, you know, when I was a kid, when I was in school, you'd go to the library to research, you'd go to the library to do a paper, you know, there'd be the old newspapers on microfiche. Do you think of like new media as being a part of the library collection?

16:24Like what about podcasts? What about videos? We're collecting born digital material. We have to, right? Any one of the formats that they want. I mean, look, that's all amazing and great. Let's just give a nod for a minute to the good old fashioned book. You buy a book, you put it in, yes, a temperature and humidity controlled environment and come back two or 300 years later and you can read it. By the way, when I was doing some research on how does one have a digitally rich medium that has the most longevity, acid free paper in a non-human environment with some oxygen of the thing is about as good as it gets.

17:03And it might be now you could digitally encode like shining a laser through a diamond or something and get it at a bitstream, maybe? No, by contrast, when I got here, one of the things I got said is we have a big problem. I'm like, tell me, that's what I'm here for. At the Library for Performing Arts, which just won a Tony Award, is the greatest collection of theater, dance, and music. and they said you should know that a lot of that collection is you know ancient tape right and it is turning literally to vinegar i said give me a jar of vinegar to take on fundraising they said no no before we get to some of our specific questions around ai in the library how are you thinking about it how are you approaching it what does it mean because obviously we're coming into the ai generation obviously the mission is how do we elevate the human mind and human beings as much possible.

17:56How are you and how is the library thinking about that? There's no denying its power. People think libraries are book depositories and lenders. That's just a great tool for sharing information that is trustworthy as a basis for people learning, creating, finding out about each other, living with each other, you know, having a life. So a tool of this power is a great thing, And we want to use it, but we don't want to sacrifice our values in the process. We recognize its power. We also recognize its constraints. And we are used to standing in the social position of standing up for values when the markets by themselves do not.

18:43That's sort of our MO and always has been, even if we weren't always explicit about it. So how can we use AI and related technology to serve that aspirational goal of every book being available to everyone on the planet for free and respecting intellectual property along the way, right? How do we use AI to help people find what they're not even sure they're looking for in that 57 million books? Or if we combine with other libraries, you know, God knows how many books, right? How do we do that in a way that is trustworthy? One of the things that disturbs me about where we're going here is the market doesn't seem to be as committed as we are to this is true information or as close to true as we can imagine it to be or certify it for being.

19:46enough processes have gone into it, not just to create more, but believable. Because the world we live in is getting less and less believable, right? Our connection to what I would describe as reality is diminishing when we need it more. So if there's a way we can use AI and its related tools to do more of what we're in the business to do and do it truthfully, do it so that you can trust it, right? Remember, trust is where I started here. That's our biggest asset, perhaps more than the books. I need to make sure we don't lose that, that asset and use it to do more so that when you are faced with 57 million books, you can find what you need, right?

20:38You can find stuff we didn't even know was there. Literally, we have stuff in the vaults here that we've never looked at, right? It's incredible, but it's true. It's a lot of stuff. The power of AI for helping us with that, with the sort, is amazing. But it has to be, in our view, sort of quality material so that we can trust it. Or we'll never trust each other. Well, you said that, you know, the New York Public Library is one of the most trusted institutions. And so with that, you have a lot of capital. to be able to sort of steer, might it be public opinion, to steer sort of the AI debate. Like, where do you see AI helping you serve your mission?

21:20And that might be helping you serve books. It might be offering AI classes. It might be giving New Yorkers free chat GPT. We're open to discussing all that. We are committed to teaching the public about AI. We'll come to the inequity questions embedded in that, right, which we recognize and want to address. We need to teach ourselves about AI. I'll just be honest, right? We're all learning. We need to use it for sort and discovery because we have just too much stuff for the human brain to take in. Doing decimal helped, but, you know, there's more to go. We need it to provide answers to people that are helpful.

22:01The gains of that are extraordinary. But we also need to make sure that we can rely on it, that the scale of it and the speed of it doesn't overwhelm the value of it. So we can do all that. We have to be really careful here. When Google started scanning books, the Google Book Project, it sounded like nirvana. We signed deals which, forgive me, sort of gave the ship away, right? Here we are in a somewhat comparable moment. but we don't want to do that. Not just for our own interests, but for society's interest and for the industry's interest. You know, the lack of concern about hallucinations, using that phrase, is not good for AI in the long term, let alone society or people.

22:49So we're going to stand up against that and say, no, you need to use certified information that's in our collections as the basis that you are pulling from and drawing from so that it is more trustworthy. One of the things that seems to be happening at the moment in AI is, you know, it's obviously growing. That's amazing. It's also producing more hallucinations. I'll give you two other examples that sort of stuck in my head. About a year, year and a half ago, I start getting reports of people showing up at the reference desk in the majestic Rosemain reading room upstairs asking for citations that did not exist.

23:35I was like, oh, that's bad, right? That's a bad sign. Or most recently, I saw the major newspapers in, I want to say, Chicago and Philadelphia published a sort of a reading list of books by famous authors that were made up books. not good, right? Google Gemini here to provide some context. In May 2025, a summer reading list that was largely fabricated by AI appeared in a syndicated insert section in major newspapers, including the Chicago Sun-Times and the Philadelphia Inquirer. A freelance writer named Marco Buscaglia later took responsibility for the piece, admitting that he had used artificial intelligence to help generate the list but had failed to verify the information.

24:20This resulted in the inclusion of numerous phantom books, plausible-sounding but entirely fabricated titles attributed to well-known authors. The incident has sparked widespread discussion about the role of AI in content creation and the critical importance of human oversight and fact-checking. I do think that summer reading list set back AI. People were so, rightly so, furious about it. Well, first they were so believing, which is says something about the power of AI. It's scale, it's speed. It's so impressive. I remember the first time I played with that, I was like, oh my God, the world has completely changed.

25:00And I personally, I happen to believe that reading and writing are core to thinking. And that thinking is the essential ingredient of morality, empathy, all of that. Not only is it so fast and so big and so impressive, because it's so impressive, it has the possibility of false trust. Simply because it's presented, the answer is right here.

25:37I do think you highlight exactly one of the things that's very important, which is understanding where this is trust, where we can build community or reliability upon it. Now, I do think just to speak of a little bit on the side of industry, being the, you know, maybe the closest to the industrialist amongst the three of us, there is some natural incentives. um it's part of all the like the chain of thought deep research citation you know people want accurate information so some players in industry are doing a lot for that part of what we're entering into is now in a much more intensified way than the internet ai will become a basis for our epistemology collectively as a society so when we're crafting this tool what is our common epistemology.

26:27And obviously that's a deeply difficult question. So is truth, by the way. Exactly. Isn't it just the politics of it, but it's also this question around like, well, okay, so part of what most people have difficulty understanding is that truth is an evolving theory of the world. Yes, but we have ideas about how to check it. We have processes to check it. we shouldn't chuck those out the door, right? But I will say a year or so ago, we got, I won't say who, but we got a bunch of requests from AI folks who basically came to us and said, okay, here's the deal. We'll pay you. But when they said, and we want to scan all your books and ingest them.

27:11I'm like, only if I know what you're going to do with them, what checks and balances are there going to be? And I remember saying this, and I think this was a little forward thinking. If you can help me create an AI based on the books, not on the large language and all the, you know, stuff on the internet that gets swooped into that, then they basically walked away. Right. And, you know, I'm hoping that conversation will pick up. We can help each other. And we're in a really good position because we're not a governmental entity, right. And we're here for the public. We want that. There are definitely some interesting partnerships.

27:55And, you know, I think there will be a number of different AI companies, AI projects. it'll be is how do we orient around trust? And as I was saying earlier, I think it's a good thing, even though difficult, to bring this collective epistemology question open. Because the reason why the epistemology question is challenging is not just because of the conflicting human interests and value interests and political interests, but also because given it's a theory of the world, and I completely agree with you, we have a variety of truth-seeking mechanisms as part of of actually how we've made the modern world as great as it is.

Read the full transcript

28:31And, you know, I also had the chuckle with the, you know, the modern health and human services, you know, pushing out the AI thing without cross-checking it and having, you know, bad citations in it. But like, okay, how do we use this to make our search for truth both individually and our discourse around it together, collectively, better? Truth is a really valuable asset. People trust us for good reason to sort of look out for society's interest in this ecosystem. And that's what we're going to do. We're going to use it to help us do our job better. We're going to say it has to look like this.

29:14It can't do this. We're in the business of, you know, veracity of things you can check. Right. So let's check. I think there is a real possibility here that we're all warming up to, which is how it should be. I don't know if you use this phrase, but the phrase that occurs to me is for public, not for profit. Right. And that's the awesome mission. That's us. So I want to learn from the past a little bit. You know, AI might be different than previous technologies. It might be the most transformative force for good. It might have more downsides than other technologies. I think it's still be written.

29:48But can you talk about sort of the past decade? What were those technological transformations that the library has gone through that you think had the biggest impact on New Yorkers and served New Yorkers the best? What was it? Yeah, we've always been here with books. The format has changed. You know, I did some research when I got here and was a little shocked by the numbers of people who didn't have broadband at home. Everything we're talking about here doesn't work for those people. So I started giving speeches about millions of New Yorkers left in the digital divide. And I got sort of no response, as far as I could tell.

30:24Then one day, I'm in the Bronx visiting a branch, which I love to do. Gets me out of my baronial office. And it's just after closing time on a beautiful day. And there's a kid sitting on the stoop in front of the branch. And I say, what are you doing? He says, I'm doing my math homework. It's online. It's a sign. I say, that's fabulous. Why are you sitting here? He has like the oldest laptop I think I've ever seen. And he says, because we can't afford broadband at home. So I sit here and get a bleed out of the door to do my homework, at which point my head exploded. I come back, I do some more research.

31:07We don't even know how many people in New York don't have the most basic thing we all carry in our pocket. This was before the pandemic. Now imagine in a pandemic, school, work, friends, life, right? And here we are. And I'll just say in the pandemic, I talked to some of the CEOs of some of the biggest providers in the country. And basically the answer I got was, oh no, we have plans that are cheap enough. I said, they can't afford them. He said, that's because they're making the wrong life choices. Again, my head wanted to explode. Well, so I'm just really excited that a seven-year-old is making the wrong life choices because he's controlling the broadband access in his, you know, okay.

31:54So I said, tell me how we can beam broadband out of our branches into every neighborhood and create what I think we should have, which is a utility level of free broadband. Not the bells and whistles, but enough, right? We tried with Citizens Band Radio. That looked promising. along the way we discovered there are other ways to do this we're starting to do it we just signed a deal with new york state for giving 2 000 low-income folks you know free broadband you know as a way to figure out how do we build this these are people who don't have money right ai is not going to work for them if they don't have access to it and what is the current like what's because like one of the things would be awesome we're still working okay got it because because that would be a good thing to replicate.

32:43I remember when I first started down this path, I said, I got some advice from folks in Washington and they said, nobody really cares about you crazy people in New York. But if you can do a pilot, which we did, with Kansas and Maine, we're going to pay attention. So that's what we did. It's delightful you've done so. And it's delightful that the drive is to do that. I mean, it is an obvious thing for, I think, every single city in the world. and inclusive, of course, very much America, to say you should do broadband coverage through the city. Because the city, it's cheap, it's easy. Well, masonry makes it a little complicated.

33:24I mean, in rural, you don't have to get through concrete, right? Yeah, exactly. But also, look, you could do a Wi-Fi mesh network, not horrifically expensively. Right, exactly. Why don't we? Yeah. And if you literally said, we want to change economic outcomes, And if you say we want to leverage what is coming in an AI universe, because I do think that AI is, because it builds upon books, the best educational technology thus far created in human history. That goes back to South Africa for me. And it's all about this institution. But there's some things we need to make sure we get right. Can I ask you one other?

34:02Yeah. But human costs here. Leaving aside the sustainability costs, I'm trusting that we're going to find solutions to that. You know, when the Industrial Revolution came and said, oh, no one's ever going to work again, and that clearly isn't what happened. This does feel slightly different. This thing at least looks like or feels like it can think, which is sort of our last, what we always said was our last respite, right? It does feel to me already that the job market, you would know something about this, has changed, if not constrained. I see people, people I love looking for work. And I think, you know, partly there's uncertainty about the state of the world and the politics and the economy, tariffs, whatever, I get that.

34:54But I think employers are saying, you know, we may not need a person to do this anymore. And let's hold off, right? I think there is a potential real cost to human beings in this. I'm not a futurist, so I can't sort of say, what are people going to do instead? How's the economy going to actually work? But I'd love to hear your thoughts, if you don't mind, on, you know, sort of why I shouldn't be losing sleep on that one. There are some unfortunately guaranteed human costs. For sure, every universe, including some very, very good ones, there'll be a lot of job transformation. And in that job transformation, this job goes away, this job gets created, et cetera, et cetera.

35:38And there's dislocation along the way. And there's dislocation along the way. And the dislocation is not, like job one decorated on Monday, job two was not created on Tuesday. So there's like dynamic processing on that. And part of the reason I call AI the cognitive of industrial revolution is because not just for the positive upside, we do not have anywhere near this modern world. We don't have a robust middle class. We do not have, you know, education, the things that lead to the public libraries and all the rest without the industrial revolution. I think we will have similarly upsides from the cognitive industrial revolution is very, very likely, but the transition cost will also be real and how do we navigate that well?

36:19So I think there's a set of guaranteed human costs in doing it. Now, the problem is, is you get to a probability assessment of, well, will it be that there's just a lot less work because a lot more of this work will be done that way? And even as the jobs, the economy grows, a disproportionate of its grab is taken over by essentially computation and automatic processes and we have to sort that out. There's a possibility. And then we'd have to sort that out. There's also a possibility, part of what, you know, in my, you know, this year's book, Super Agency, is that actually, in fact, what it does is it's the same way you say, I'm a professional and I don't use a computer.

37:02You're like, well, actually, in fact, all professionals use computers these days. I'm a professional, I don't use a smartphone. It's like all of them use. I'm a professional, I don't use search engines. No, all professionals use search engines. It may be the, I'm a professional, I don't use AI. It's like, no, no, all professionals use AI. and even a librarian deployed will be deployed with multiple AI agents. And the mistake that usually then people then go is they say, oh, if you have this productivity enhancement, there's a fixed amount of work. And so there'll be less work. It's like, no, no, usually work grows.

37:30We know that's not the case either. Work grows. Transition's still very much there. I tend to think that the right way to think about this is to be intelligently optimistic, not blindly optimistic to say, yes, it's likely that there will be a lot of stuff there. We just need to navigate to it. If we end up in a Star Trek universe, in which case we just have a lot of work being done by robots, by AIs, then the only real question is to make sure that that's, it's not like a small tier of people, like Elysium, the movie, you know, like having a elite life with everyone else. And the equitable part, can I jump in just for a second?

38:08It worries me a great deal. There's two examples I'll give of that. One, forgive me, capital versus labor. This is a form of capital, not labor. And by definition, that means it can skew further, right? And we've already seen more skewing than, in some ways, more than ever in history already, right? Well, not ever in history, post-industrial revolution. Pre, it was medieval times. Remember kings, peasants, et cetera? All right, all right, all right. A little hyperbole. All right. The other is, I think about education, right? Very powerful tool for education. Makes it, you know, can make it bespoke, can make it responsive, can the teacher remembers everything, personalized feedback, you know, more resources, obviously.

38:55as a former college professor, I see technology improvements to education that have many of those plus sides. The other thing I notice is they tend to be given to less privileged kids as a way to fill in what the system isn't doing for them. And my kids and your kids are going to places that look remarkably like they always have, right? Which tells me, revealed preferences, they're better. Or we think they're better, right? So I worry that the technology can, not just in economic terms, but in educational terms, increase the gap. Well, I'll jump in on the education piece because we have Mackenzie Price on the Possible Pod in a few weeks, and she is running a private school, Alpha Schools, that is all about AI first.

39:50And it's an expensive private school that parents pay$50 ,000 a year to send their kids to. And, you know, kids are using... Wait, where is it? In the Valley? Austin, actually. Kids are using AI two hours a day to learn, et cetera, et cetera. The rest of the day, they're not. They're cooking, they're debating, they're learning about ethics and morals, et cetera, et cetera. And this is what the fancy people have chosen. So my point of view would be the public education system is already broken. I'm lucky that I have a good public school in my neighborhood. And my dream is that AI will be the thing that breaks it open so that we can change our public schools.

40:27And so there's one way where, right, you could go a bad way and have, you know, okay, the poorest among us just get to sit in front of an AI tutor for 10 hours a day and don't get anywhere. Or you have a reimagination so that AI is forcing us to rethink this and get to a better place. Because the power of personalized feedback to be the equivalent of a tutor for a poor kid is truly remarkable. We just need to make sure it's not the only thing that they're getting. And so I would like thinking about New York City, everyone has access to a public school and everyone has access to a public library.

41:05And so I would sort of ask, like, what is the way that we can make sure that the poorest kids, the poorest adults, low-income New Yorkers get the best of AI from the library? I would join that conversation with pleasure. When I got here meeting lots of people, meeting technology people for the first time in my life, well, one of the people I went to see was the chancellor of schools, of course. Biggest library, biggest school system. As it turns out, Dennis Walcott, who's amazing, is now the president of the Queens Public Library. So, you know, lots of connections. And I said to him, you know, we're these two systems.

41:47We've never really talked to each other. We just sort of assumed we were doing what we needed. I said, what do you need? He said, well, I can't afford the libraries in my schools anymore. And I said, well, that's funny. I've got 50 million books sitting on shelves most of the time. Seems like there's just a disconnect problem here. And it turned out we became the libraries in most of the public schools, all three systems, which Dennis now runs one of. And it was one of these like, oh, of course, like we just forgot. Similarly, when it comes to an AI, I'd love to play, right? Right. And, you know, it's an amazing school system and it's an amazing library system and AI is amazing.

42:33Why shouldn't we figure it out? Right. Can libraries be these AI centers? Well, we're going to start by training the public and training ourselves, et cetera. But that's just the start, right? But by the way, how do you best train the public at scale? Use AI, right? Number one. Number two, look at the hopeful nudge on this. As long as it's reliable. I'll stick with that. No, no, but actually, in fact, learning both make it more reliable, but also learning to approach the outputs just like you approach any other output, including the output of a book with a critical mind, is really important. Like just because it's in a book doesn't make it true, right?

43:09For example, I use AI very repetitively for is like if I think of something as an argument, I will frequently put an AI and say, give me the counter argument. We're all going to learn how to prompt. Exactly. But also, it's a little bit of what surfaces and affordance you give. For example, if part of how you're training AIs to do this is to say, make sure the person's engaging in critical thinking. So if they make an argument, like flag the counter argument as an example, those are all doable in this. Also, think about the political effect of that, right? I mean, that's fabulous. I was once chairing some national commission, and I said, one of the things we should say is, you know, we should force people to confront alternative views.

43:54And the social science I got hit with was, oh, no, that will make people back in more where they are. I was like, that can't be true, right? Like, if we believe that to be true, we are toast. If we can get AI to help us with this, because we're not doing it for ourselves. So I think we can. The question is, will we? But because this is such an important institution, one of the things I want to express as a moment of optimism around this is wealthy people have always been able to afford tutors for their children. One of my hopes for AI is tutors for every child. It's a little like people used to say to me, oh, the evidence says smaller classrooms doesn't help.

44:33And I said, that's funny. Why are you paying$60 ,000 a year for smaller classrooms? I don't believe you. No, absolutely. I mean, I think like my main worry, it's funny, I had a friend who was saying, hey, I want to get into AI. I'm a social impact guy. I want to make sure that AI doesn't harm people. And I was like, that's great. I also want AI not to harm people. And I think of all the good people get together and all they focus on is having AI not harm us. We're going to miss out on all of these incredible things that we can do. And so, you know, I'm so excited for the New York Public Library to be thinking about, right, both how can we save truth and the same New Yorkers that I want to get broadband for.

45:11Or tutoring, individualized tutoring, right?

45:17So, Rapid Fire, is there a movie, song, or book that fills you with optimism for the future? I had a chance, finally, way too long, to read Robert Caro's Power Broker, which my father told me was the best book on politics he'd ever written. And probably that's why I didn't read it right for a few decades. And that may surprise you in terms of optimism. It shows that things can get done at big scale, which we need, need more of. And it shows that the people will contain it when it goes off. It took a long time. Maybe that's not everyone's interpretation of what happened, But both of those are in that narrative.

46:08And I thought that was amazing. Amazing. And I learned a few tricks. Well, I think the New York Public Library is a testament to just that idea. Like if some politician came and said, hey, New York, we're going to build hundreds of buildings, staff them every day from nine to nine. We're going to spend millions, hundreds of millions of dollars on physical items to put in them. Everyone would say, you're effing crazy. We're never going to do that. Oh, immigrants come to this town. We're the first stop that they come to. And they're like, you do what? You like give this for free? Like, you know, let alone, you know, this building and, you know, what it has to offer, right?

46:46One other book that I haven't gotten out of my head for now a couple of years, it's not generally thought of as optimistic, but I think of it as, is Ministry for the Future. Amazing novel about what climate change is moving us towards. I don't want to ruin this, but it's incredible because of the creativity of seeing ahead, you know, in a very real way and a very dramatic way. And we figure it out. We agree. We've had Stan on the podcast. Oh, good. Good. OK. So amazing. I'm jealous. Tell him I said I. This could be personal, professional. What is a question you wish people would ask you more often?

47:29Sure. How can I help the library? Everyone should ask that. I love that. For all of our listeners, fantastic. So where do you see progress or momentum outside of your industry that inspires you? The improvements in terms of health is the most obvious one, I would say. I mean, in creativity and the arts is just, you know, what's happening, which often happens in hard moments. I lived through South Africa in a revolution. I know, I think we're going to have breakthroughs in the energy world, a la Robinson's book, transportation, all that. I mean, we already have it. We're going to have more of it.

48:13I think human capacity for ingenuity, for finding solutions, for being creative is never going to be replaced by any machine and is our most valuable resource. And we have to keep it coming. Yeah. Nurture it. This is my most fundamental belief, right? It's why I'm so happy to work in a library. It's what I learned in South Africa. It's everywhere. Don't count it out. You'll be wrong if you do. I love that. Can you leave us with a final thought on what is possible to achieve over the next 15 years if everything breaks humanity's way and what's the first step to get there? You know, things like health and energy, I mean, things that if we can like find a solution, a solution will be transformative of our global experience.

49:10There's a lot of people on this sweet little planet, right? We gotta make sure that everybody's okay. The way to unleash that is to unleash the human capacity. interaction with whatever tools we can perfect to help us. We need everybody doing that. I love it. Thank you so much. A complete pleasure. Thank you for hosting us here at the library. Really, my pleasure. This was more fun than I expected.

49:41Possible is produced by Wonder Media Network. It's hosted by Ari Finger and me, Reid Hoffman. Our showrunner is Sean Young. Possible is produced by Katie Sanders, Edie Allard, Tanasi Dilos, Sarah Schleid, Vanessa Handy, Alia Yates, Paloma Moreno-Gimenez, and Mulea Agudelo. Jenny Kaplan is our executive producer and editor. Special thanks to Surya Yalomanchili, Saida Sepieva, Ian Alice, Greg Beato, Parth Patil, and Ben Rallis. And a big thanks to Matt Kirby, Jennifer Firmino, Connor Goodwin, Claudia McKinney, Elizabeth Gilroy, Valerie Visconti, Johnny Sang, Renee Gonzalez, and Johairi Dela Cruz.

From the publisher

What can public libraries teach us about finding and sharing trustworthy information in the age of AI? How can—and should—AI build upon the technology that is books? This week, Reid and Aria head to the New York Public Library in Manhattan for a conversation with the storied institution’s president and CEO, Tony Marx. They explore the function of libraries in the age of AI, debate the future of work, and discuss what it looks like for civic institutions to embrace transformative technology without sacrificing core values. 

For more info on the podcast and transcripts of all the episodes, visit https://www.possible.fm/podcast/ 

Topics:

2:54 - Hellos and intros

3:23 - Tony’s studies and work in South Africa

5:25 - Taking a job as president of Amherst College

8:43 - Eliminating library fines and fees

9:23 - Visiting the Inwood Library

13:30 - Determining what gets digitized 

16:20 - New media 

17:42 -  AI’s function for libraries 

23:08 - Navigating hallucinations

27:57 -  AI becoming our epistemology

29:36 -  Technological advancements at the NYPL

33:54 - Future of work and human costs of AI

39:41 - Can AI close the educational gap?

42:43 - Trust, reliability, and AI

45:18 - Rapid-fire questions

Select mentions: 

Possible | Kim Stanley Robinson on the future of civilization: https://www.youtube.com/watch?v=B6OlJbPDpyc 

Possible is an award-winning podcast that sketches out the brightest version of the future—and what it will take to get there. Most of all, it asks: what if, in the future, everything breaks humanity's way? Tune in for grounded and speculative takes on how technology—and, in particular, AI—is inspiring change and transforming the future. Hosted by Reid Hoffman and Aria Finger, each episode features an interview with an ambitious builder or deep thinker on a topic, from art to geopolitics and from healthcare to education. These conversations also showcase another kind of guest: AI. Each episode seeks to enhance and advance our discussion about what humanity could possibly get right if we leverage technology—and our collective effort—effectively.

More from Possible

All 83 episodes
Tony Marx on the Future of Libraries, AI, and Trust Possible · 52 min
Listen in VO