← New Books Network
Loading summary
Transcript53 lines
- [00:00]
Emily Bender
K Pop demon hunters, Saja Boy's breakfast meal and Hunt Trick's meal have just dropped at McDonald's. They're calling this a battle for the fans. What do you say to that, Rumi? It's not a battle. So glad the Saja Boys could take
- [00:12]
Jeffrey Hurley Humera
breakfast and give our meal the rest of the day.
- [00:14]
NBN Announcer
It is an honor to share.
- [00:16]
Alex Hannah
No, it's our honor. It is our larger honor. No, really, stop.
- [00:22]
Emily Bender
You can really feel the respect in this battle. Pick a meal to pick a side
- [00:27]
Jeffrey Hurley Humera
Ba da ba ba ba and participate
- [00:29]
NBN Announcer
in McDonald's while supplies last. Hey NBN listeners. We're running our 2026 New Books Network Audience Survey and we'd love just a few minutes of your time. NBN has been bringing you in depth conversations with authors and scholars for over 15 years. We haven't done a comprehensive audience survey since 2022, and a lot has changed since then. It's time to hear from you again. Here's why we're asking. We want to understand who's listening, what subjects and podcasts you love most, and where you'd like to see us grow. Your responses help us tell NBN's story to the publishers, libraries and institutions we partner with, when we can show that our listeners are serious readers, lifelong learners, and heavy library users. It opens doors to new partnerships, better resources, and ultimately a stronger NBN for everyone. And one more thing. If you leave your email address at the end of the survey, you'll be entered to win a $100 gift card to bookshop.org, a chance to stock up on books while supporting independent bookstores at the same time. The survey takes just five minutes. Your answers are confidential and your email will never be shared. Head to newbooksnetwork.com to take the survey today. We really appreciate your support. Now go take the survey.
- [01:45]
Alex Hannah
Welcome to the New Books Network.
- [01:49]
Jeffrey Hurley Humera
So greetings from my way of Puerto Rico. My name is Jeffrey Hurley Humera. I'm a professor in the Department of Human Humanities at the Universidad de Puerto Rico, Mayaguez. This podcast, Nuevos Orizontes, discusses cultural studies, art, thought, technology, literature, philosophy, and decolonial themes with a Caribbean axis anchored here in Mayagues. It is sponsored by the Humanities Department and in part by the Tigo Foundation. Today's episode is an interview with Alex Hannah, Director of Research at the Distributed AI Research Institute and Emily Bender, professor at the University of Washington. They are co Authors of the AI how to Fight Big Tech's Hype and Create the Future We Want 2025 Harper thank you so much, Alex and Emily, for connecting with us today.
- [02:29]
Alex Hannah
Thank you for having us.
- [02:30]
Emily Bender
It's a pleasure.
- [02:32]
Jeffrey Hurley Humera
All right, great. Before we start our conversation, I also wanted to introduce our co host. Alex Rivera. Cartagena is an instructor of philosophy at the UPRM and also a UPRM graduate. And like myself, before coming over to the humanities, he studied computer science.
- [02:46]
Alex Rivera Cartagena
So yeah, it is a pleasure to be here and it is so wonderful having both of you in this podcast. It is so wonderful for me to have this opportunity.
- [03:01]
Emily Bender
That is very kind. It's really delightful to be in conversation with you.
- [03:04]
Alex Hannah
Yeah, likewise. Thank you.
- [03:06]
Jeffrey Hurley Humera
Thank you guys so much for taking part in this conversation. And by the way, this is the first of two podcast episodes that we're doing on the AI conversation. The second, in Spanish, will appear on the Nubucks Network in Espanol. Today's episode, we're taking a critical look at one of the most powerful and contested technologies of our time, artificial intelligence. While AI is often presented as an inevitable, revolutionary, unquestionably beneficial, in the AI con, Alex and Emily argue that what is being marketed as transformative intelligence is in many cases a carefully constructed myth, one that obscures issues of labor exploitation, environmental cost, institutional complicity and further concentration of quantities corporate power. In the AI con, Alex and Emily examine how these systems reproduce social biases and make a persuasive case for why universities and public institutions should should be especially attentive to the ethical implications involved. Rather than accepting the dominant story of innovation and inevitability, the AI con invites us to pause and ask harder questions. Who benefits from the current AI boom? What values are being advanced or undermined? And what would a more accountable and socially responsible technological future look like? But one of the most important aspects of my reading of this book is how this impacts community on human to human interaction, on consciousness, and on the ways knowledge itself is formed through those person to person encounters. These are pillars of what universities ostensibly seek to cultivate and to sustain. The four of us on this call are fortunate enough to have used pencils, to have gone to libraries and sought guidance from people there. We learned to read, write and create knowledge without algorithmic mediation. We are increasingly in the minority in having those formative experiences. And the question now is not simply whether the new tools are efficient, but what happens to shared intellectual life, to mentorship, to attention during the slow destructive process through which automated systems become the default co conditions of learning. I am thankful for books like this one, and of course, for the people who write them. Before I pass the Mic to Alex, who has our first question. I wanted to say thank you again for your generosity with your time. Thank you. And also for your commitment to ideas like the ones in this book.
- [05:10]
Alex Hannah
Thank you. That was a really generous introduction of the book. So thank you for providing that.
- [05:17]
Jeffrey Hurley Humera
You're very welcome.
- [05:19]
Alex Rivera Cartagena
To begin. I wanted to say both of you, thank you again. You pointed a problem that it is a concurrent and important problem in the AI discussion. It is this hype, the hype that it is around about this technology that has so much promises, but in reality are so empty. I must admit that I was once a victim of this hype. When the model ChatGPT 2.0 first came out, I believe many of its promises. I was impressed with the mere change idea of asking an AI questions and being capable of generating coherent text with answers that seem convincingly true functionally. With time, I realized that everything related to AI is an act and that there is nothing intelligent about artificial intelligence. It is only that, an artificial emptiness. It is filled with erroneous information and means, translated texts. And I ask myself, do we really want this technology to replace us? Unfortunately, we live in a society that is closely related on technology, which is leading us down a path that may seem inevitable and not be easy to fix, one where we want humans, being human beings, to be machines and the machines to be like human beings. But now I wanted to ask, tell us a little about your biography, your instances, training and experiences and how they have crystallized in this book.
- [07:17]
Emily Bender
Thank you. And thank you also for those lovely reflections on the book. Speaking for my side of the story, my training is in linguistics, and you can understand linguistics as the study of how language works and how we work with language. And I have been running our professional master's program in Computational Linguistics at the University of Washington since 2005. And so I came into this moment of AI hype from the field of computational linguistics that was already experiencing it a little bit earlier, before OpenAI sort of imposed their product on the whole world and coming at it with some. Some disciplinary knowledge that I thought was really important to share with the world about how language models work, how we know that they don't understand, but why it is that they seem to understand what we're saying. And then another part of my story is that through getting involved in looking at the possible negative societal impacts of language technology and being part of conversations around that online, I got to meet Alex and start working with her. And then we started having our own podcast. And the podcast led to the book. But I'LL let Alex pick it up from here and tell her story too.
- [08:30]
Alex Hannah
Yeah. And just to give you a little bit of my background, I'm a sociologist and originally focused quite a bit on social movements and how they use technology. And so in my experience, I had been working with different machine learning tools for computational social science analysis. The more I got enmeshed in this, the more I realized how much different institutions are interested in these types of tools and how they could be used for things like surveillance and harms to marginalized communities. And so got much more interested in that. In particular, very interested in the data sets that are constructed around these technologies, as well as the labor practices that started the conversation around these benchmark data sets that are used for evaluation in those conversations online, Emily got involved and we wrote several papers with a few different authors, including debraji Remy Denton and Amanda Lynn Palata, and then found an
- [09:39]
Jeffrey Hurley Humera
article,
- [09:41]
Alex Hannah
had a group chat talking about a lot of the different hype around these different models and really wanted to make fun of one because of how bad it was. And so I said, oh, why don't we try doing this on a. Or rather Emily asked, does anyone want to do this with me? And I said, oh, yeah, why not? Let me try to set up a twitch stream, because I heard that's something that people do. And we set it up, had a lot of people come. It took us three episodes to get through this very long article. And then from there we formed an accidental podcast ourselves called Mystery AI Hive Theater 3000. And then after we had been a few episodes in actually quite. Quite a lot of episodes in, we were approached with the idea of writing a book. Initially, Emily was. And Emily brought me on board. And the rest is history.
- [10:33]
Emily Bender
Yeah. Although Alex left out one important detail. So we started interacting online and actually working together on papers in 2020. And so I think three academic papers, a handful of op eds, 50 podcast episodes, a whole book. Later in March of 2025, we finally met in person.
- [10:50]
Jeffrey Hurley Humera
Wow.
- [10:50]
NBN Announcer
Incredible.
- [10:51]
Jeffrey Hurley Humera
Wonderful. It's really exciting for me to have both of you guys on as far as linguistics and that really language is a central part of what we do in Puerto Rico, but we're the only Spanish language Land Grant University. And we are really work through Spanish and into English into the kind of the technical fields. And I feel like we have a really unique kind of linguistic context here. And also Alex being, you know, very interested in social justice and resistance and solidarity. And in fact, I wanted to kind of parlay a comment that Emily just made that about how AI has imposed a product on the whole world. In a previous podcast episode, Alex and I discussed Shannon Valour's book the AI Mirror. And Alex said something that has stuck with me. And he said that it's about AI interference, which is constant, relentless and incredibly damaging. And he was not speaking about isolated classroom incidents or occasional misuse, but a broader atmosphere, kind of an ambient presence that subtly reshapes how students approach reading, writing and even thinking. And this interference is not dramatic, it is incremental and it is subtle. And I think it enters at all different levels in the classroom of how we summarize, how we brainstorm, how we create drafts, the idea that over time, the complexity that students would have as a skill set, patience, tolerance with, for ambiguity, confidence in one's own voice. I feel like all of these things are being lost. And we have a colleague here who unfortunately couldn't be on the call today, Hector Heike, who published a book last year called Elohio las Circanias or Eulogy to Closeness. And I think what concerns me most about how this constant mediation alters the relationship between student and student, student teacher, even between a person and their own thoughts and their own thinking process. And once this kind of reflection is outsourced, if a language is always pre shaped, if interruption and interpretation is pre digested, what happens to what we're supposed to do in the university? At what point does this become, as Hector Hake, as my colleague says, substitution? And my question comes to that is when we think about that, about what's happening, and we have some excellent people like you guys and also here on our campus thinking about this and you guys have any reflections about what is the capacity for universities to confront this? How can we conceptualize a regulatory action to address what's happening?
- [13:07]
Emily Bender
Yeah, this is something that is super frustrating and I really appreciate how you articulated it and how this idea that these systems could somehow replace the work that we're doing, I think really flattens what it is to be a student and to be a teacher into the words that we exchange rather than the relationships that we build, the thinking together that we do. And it is extremely frustrating, speaking from another land grant university, to watch my institution, at least the upper administration parts of it, getting super excited about this technology and saying we have to reshape everything we're doing to, you know, prepare people for, prepare our students for the AI workplace or whatever, when it's all just marketing and we still have the same responsibilities that we have had to students to, you know, think with them and Help them hone their own voices in whatever languages they're working in and help them also hone their critical thinking skills, their ability to make sense of the information environment and their disciplinary expertise. And to watch all of these people say, oh, yes, well, that looks good coming out of ChatGPT, so that's better is really frustrating. And just the cherry on top is I recently discovered there's a webpage from the graduate school here at the University of Washington that is talking about ethical and responsible use of, in quotes, AI in research and writing. And it includes a whole bunch of stuff that was extruded from Copilot and ChatGPT. And it's like, do these people actually not want to do their jobs? Like, why are you here if you aren't going to be in the business of engaging with students and helping them grow as scholars? I don't think I answered your question. I think I just gave you back a bunch of frustration. But I do think that the answer comes in faculty organizing and faculty working together to, you know, protect our workplace and protect sort of the value of what it is that what we do. That we do.
- [15:07]
Alex Hannah
Yeah. And the way you phrase that, I think was really helpful, Jeff, because I think that what you said, I think is really important. I think it's often not about, you know, the tech specifically. It's about the kind of educational environment that the kind of the idea and the idea of the tech enables. So the I. The kinds of things and I. And one thing I think we do well in the book is how I think we situate trends that have been existing in the university and in the workplace and in medicine within a historical context and in the university, especially the American university. We are in a prime neoliberal era. We are in this era of massive debt financing, of rising tuitions, of the reduction of state funds to fund ostensibly public universities, casualization of academic labor and modularization of curriculum to make it very easy for anyone, seasonal or adjunct labor, to teach it. That is coming at the harm not only to the instructors who have to take these terrible working conditions, but also to students because now they are, they are engaged in this very product oriented, consumer oriented relationship. And so it's really awful because it doesn't really think about. It doesn't really make us think about process, about what the educational process should and ought to be, that one has to struggle through writing or struggle through thinking through different, struggle through the text, struggle through reading. And, you know, it's kind of. I hear the same thing at every university I go to from instructors and writing center coaches and TAs. This is kind of the thing that is really pushed on us. I was at Duke recently and it was, I think, the first university that had its own chat GPT. They called it Duke dpt. And it was very, it was interesting because the other faculty there brought, you know, they started a humanities seminar specifically to engage in, I think that really interpretive and difficult work of I think, confronting ideas, facing how we encounter each other and through our art and culture and really registering that that is part of the thinking process rather than, as Emily so well put it, you know, just these, you know, these language type artifacts that are bereft of meaning.
- [18:06]
Jeffrey Hurley Humera
Thank you. Especially what you just said about struggling. And those are the types of words I feel. In fact, I had an argument with some administrators here about the word confusion, about how that's a valuable situation where you, you want to create confusion people, to navigate it so that people are uncomfortable. And that's how, you know, our, our experiences grow and how we are, you know, able to navigate confusion. Confusion from Latin? No, with the, the multiple integration of multiple things. I mean, those are, those are things we seek. You know, I feel that. And when we kind of default to, as Emily excellently said, this is all hinging on what the marketing says and kind of a Wendy Brown dystopia even worse than what she imagined, you know, 10 years ago.
- [18:44]
Alex Rivera Cartagena
Or another thing that I connect with this thing. It is that the chapter three of your book. It is the notion that this technology can affect the marginalized people, but also the services that we receive, like for example, medical assistance. It is fascinating because now I'm in a class that really caused a textbook that he has a notion that technology and progress as general can give us the opposite notion. That the real privilege is not using these technologies, it is avoiding it. So basically this becomes. From Ivan Eilish. It is the name of the author. He says something like the sentence, defense against the damages inflicted by development rather than access to some new satisfaction has become the most sought after privilege. And now connecting that notion to the next question, it will be. Artificial intelligence has roots in systems of power that have historically marginalized certain groups. How do you see racism and classism embodied in early development of AI? And how does that affect the biases we see in current IE systems?
- [20:38]
Alex Hannah
Yeah, so there's, I mean, there's a lot of, there's a lot to be said here and I think there's a lot to be said both in terms of how we conceptualize AI, how we conceptualize data and then how we conceptualize deployments of these systems. And so I want to talk about them all the different levels, because I think those are all pretty critical. So first off, the term AI, you know, we say in the book, AI is a marketing term. And this term has been around since 1956, when John McCarthy coined the term at the Dartmouth conference, effectively coining the term to exclude someone else. Norbert Wiener, who had been this pioneering scholar in cybernetics and didn't like him, just didn't want to invite him. And it brought that together a bunch of different things. It brought together researchers and kind of psychology and. And neuroscience kind of in. But also in. In physics and math. And so this was a general term meant to apply to a lot of different technologies, including automatic languages, which are what we'd call to get a programming languages, but also people looking at these neural networks, which are named for a specific model of how neurons in the brain were thought to work at the time that were in the 40s. That has stuck as a term. This notion of intelligence, as you may know, has a very long and eugenicist history, especially the goal of quantifying intelligence, having its roots initially with Alfred Binet and in his testing for student aptitude. That then became a test that was formalized and used to differentiate individuals in the uk initially for differentiating between people who had developmental disabilities or cognitive disabilities, and then basically labeling them for negative eugenic policies, so preventing them from procreating or marrying and enforcing sterilization, as well as positive eugenics, trying to promote the. Promote people who were deemed more. More cognitively, quote, unquote able. And so as with many things in the US that is imported and becomes. Became racialized and to differentiate between black and brown people, but also Eastern Europeans and Northern Europeans, and that is persistent throughout. So it's very easy to see now when the people involved in AI try to reach for a definition of intelligence, they must then fall back on these types and metrics of intelligence, which basically are constructed by defenders of eugenics, because that's, you know, any kind of reification or metricization of intelligence leads down that path quite quickly. So that's in the development of AI and the data sets, these data. And I think this is where tech companies like to say that there's a solution for this. There's the biases that are embedded in training data, because there is, because the language and images on the Internet are racist, because we live in racist societies and they're also classist and. And the intersection of race and class and gender. I should mention as well, that's why we see the things like the perpetuation of racialized and gendered stereotypes. It's why there's, I think, reporting that come out about displaying images based on text prompts for what we call in the book synthetic image extruding machines. So if you type in housekeeper, it will present an image of a darker skinned woman. If you type in CEO, it'll present a lighter skinned man. And so those are at the. And that's based on what's in the training data. That is something that is more palatable to tech companies. They say, well, we can kind of prevent these or we can have some kind of mitigations, but there's not really any way to remove that. That's, I mean, this is what it's presenting and these are, this is what's baked in. These are biases that are going to be there and are difficult to mitigate. And then lastly, in application itself, knowing that these are tools that are going to be applied in places where those who cannot afford social services are going to be presented with the tools that are cheap facsimiles of the types of services that those with means who are white, who are based in more affluent areas will have access to. So tutors that are, that are in person, very high touch, health care, etc. Etc. On our podcast we had a great guest who is a community based researcher at dare, Adrian Williams. She was teaching at a charter school in the Bay Area in Contra Costa county which is a very. In west Contra Costa county, which is a poor school district. I just actually got an email that they're laying off hundreds of people in the school district because of the budget shortfall. And she was teaching at a charter school and the charter school is highly technical and had many tools that were supported by the Gates foundation or the Chan Zuckerberg Initiative. And those students were getting, you know, had to type in everything that they were doing at the school in this text box. Meanwhile, you go in different places and go to private schools, go to schools where more affluent children go and there is not a computer in sight, or if there's a computer, it's over there and it's optional and there's much more hands on instruction. And so this is the kind of thing where these things are not racist because of, solely because of the bias, but also because the histories of this term of intelligence and how they are deployed. So they're deployed in this racialized class way that really is injurious. Of these communities.
- [27:39]
Jeffrey Hurley Humera
Thank you, Emily. Did you have another response for this one or.
- [27:44]
Emily Bender
I mean, I think Alex has covered it really thoroughly. But just to put one more sort of fine point on things, the very idea of intelligence is the idea that you can rank people according to one property. And that idea has eugenicist and racist roots throughout, through and through. And anytime someone says that they're computer is intelligent at some level, on some imagined scale of people, even if they're not using iq, if they're using like people by age or people by education, they are referencing that same racist eugenicist concept.
- [28:20]
Jeffrey Hurley Humera
Yes. And these all kind of also merge over into, into language, into linguistics. I remember when I was a graduate student, I went to school in Spain, and even though I studied in a program called Arte Literatura Pentagon, I mean, the art, literature and thought, I kind of fell in with a bunch of linguists who were my friends. And we used to have these unbelievable arguments about the Spanish language because. And I remember one of the things that they would always kind of, you know, side to was they would go to Google and say, look, Jeffrey, this is, this is not. This is this use of le or low or whatever and say, see, they put together engrams things to kind of prove their points. And this was in the early 2000s. I remember thinking to myself, this is nothing more than a socioeconomic measure of who has access to the Internet, who can clog up blogs. And at that time there were many more people in Iberia who had the ability to do that. So I guess that makes their language use the important one or the correct one. And this AI circumstance is scraping all of that, you know, of those things. And your book made me think a lot about Puerto Rico too, where I've lived for the last almost 20 years. And among the things that I really enjoy about life in Puerto Rico, and one of the most important things about Puerto Rican culture, I think, is the participatory aspects. If we're in a classroom or a hallway or a gas station or a Senate meeting, there is a tradition of dialogue here that is very unique and largely in relation to that culture of civic participation. Puerto Rico has among the highest voting rates in this hemisphere. And when we think about Puerto Rican, what Puerto Rican institutions can offer to other countries and specifically to the U.S. one important thing is the democratic basis for civic affairs, which in my opinion is often lacking on the mainland. And I wonder what AI what the empire is doing to all of that. And like Emily said, this imposition, it pushing it into our realities. What is this doing to the Puerto Rican cultural, linguistic and democratic institutions? And I don't, can't. You can make some arguments, I think some apologist arguments, but, but here we are in a forum about science and engineering and about AI and the humanities, arguably the most important university in Puerto Rico. And from, from your guys, sets of experiences, when you're writing and you're thinking about this topic, how is Puerto Rico perhaps more vulnerable than other communities?
- [30:19]
Emily Bender
So I mean, first of all, I think that it's important to think about vulnerabilities, but also important to think about strengths and sort of areas that should be protected. And this culture of democratic participation that you're talking about seems like something that could be a really good sort of point of resistance, but also something that will be under attack because the idea that we can automate that, that it would be somehow better because you would have more voices, you would have, you know, more people's input rather than, you know, more chances to connect and actually be face to face is something that, that I would worry about, given what you're describing. Another thing that I would worry about is that oftentimes the imposition of technology is presented as if it were philanthropy that, you know, school systems everywhere in the US And I would be shocked if this weren't also true in Puerto Rico, are usually conceived of as sites of need, of like bottomless need. We don't have enough teachers, we teachers don't have enough time. And then the tech companies can come in and maybe they're selling something or maybe they're giving it to you and it looks like it's for free. And then I'm afraid that, that the decision makers won't actually take the time to figure out whether this is really worth it or not. Because if it's being offered for free, it seems like, well, why not, right? But the thing is, it's not really free because it is, among other things, a way for these companies to gain access to data, to gain access to just like the look of having done good, right? So there's, there is something that they are taking from the school systems when they offer this for free or whatever other system it is. And so I really hope that policymakers will have the time and place to say, is this actually worth it? And if they are not given the time to make that decision, if it's rushed in some way, then that's a really good sign that this isn't free. Right? The pressure is coming. If it's a hard sell to take this free thing Then it's not free. Right. If the person offering the technology were really doing something out of just the desire to be helpful, then they should be perfectly fine with the community that that technology is being imposed on, taking their time to work through a democratic process and decide how it would really fit into their communal spaces.
- [32:46]
Alex Hannah
Yeah, I've been thinking a bit, and I don't know, and I'm unfortunately not as educated as I could be about the history of Puerto Rico and specific vulnerabilities. I was on the island in late last year. I think I mentioned this and also went on a great decolonial tour in San Juan. And what it makes me think of, just from that experience, is the ways in which these technologies often reinscribe colonial pathways. And so there's been. This metaphor has been used a few different times. I know there's the text by Nick Codry, and I'm not going to get this. I need to look up the names so I don't mispronounce them or not mispronounce. I will probably mispronounce, but not get them correctly. And then there's. I'm going to look that up real quick. And then there's the Empire of AI or Empire of AI by Kieran Howe. And so this metaphor of data colonialism is very present. And so I think thinking through that and thinking especially the way in which PR is the world's first colony and the ways in which technological colonialism also follows many of those pathways. What are the different resources or what are the different ways in which those pathways are chased? Is it, you know, is it. Is it land? Is it water? Is it a particular regulatory environment that allows for the building of things that seem like they're good deals, as Emily is gesturing to, you know, in. In the AI race that seems to be data centers, we're seeing that on the mainland quite a lot with the imposition of data centers in many different places around the country, including in places that you'd expect to be quite sympathetic to big business. But we're seeing opposition to data centers in rural Ohio, where my family lives or my mother lives, excuse me, in Texas and in Indiana, and all these different places that we don't think are necessarily kind of like places where you'd find this resistance. And it's often because there's this idea that you can kind of get away with this in this lax regulatory environment. And so, yeah, I mean, thinking about how those pathways get reinscribed, how they trace the kind of, you know, the kind of subjugation of racialized subjects, how there's specific land and water and energy grabs. And we're kind of, I mean we're kind of seeing this in PR to some degree with that, like the crypto community, I will say, like from what I know of it, which is very alarming. It's quite concerning that that kind of same pathways would be also something that AI people would want to get into as well. And just to return to the first reference of Nick Coldry and Ulysses Medgs, so they have an article in television and New media from a few years ago that's data colonialism rethinking Big Data's relation to the contemporary subject. So that was the other reference I was thinking through.
- [36:44]
Jeffrey Hurley Humera
Thank you. If I could just jump in here, Alex, before I know you've got the next question, but I wanted to just say a quick follow up. What Alex just said about. I actually said before, I meant to say this before about these ideas about racism being baked in and that there not being any visual, any difficult to mediate. Sounds a lot, sounds a lot like the responses to offline racism too from, from, from politicians. And that was, that was within political reform. That was something that I, that really kind of gave me a, a chuckle. One thing about also about Puerto Rico specifically that I wanted to, to just follow up with is that maybe, maybe since we are on an island and maybe since we have a little bit of linguistic, some healthy linguistic and political distance from the US that, that we would be able to, to hear have. I mean I, I feel like we have some certain, the students and the faculty have a certain set of decolonial sensibilities that are not really available elsewhere due to the linguistic environment and also the environment, the colonial environment at the university. And I feel like that there have been developed some specificities in the way that we create knowledge that I feel like we might have a little bit of shielding from the outside, but maybe not. But maybe the English centrism will eventually kind of overpower and just like the AI is trying to. But I would like to just one, one other thing before I pass the mic to Alex is that after Hurricane Maria, not that this is something positive, but after Hurricane Maria there was always a question that everyone would come up to when you hadn't seen anybody in a couple months and there's been no electricity for months. And the question was always, what have you been reading? Because everyone went back to, you know, the books that they had on their shelves that they'd been ignoring without, without power. Everyone was able to kind of reconnect with that in that time reminded me when I was a graduate student, I spent some time in Cuba. And in Cuba it was kind of similar that sometimes there was power, sometimes it wasn't. And there was much more of a. Of a culture of being in the street and kind of really connecting with your neighbors. And, And I felt that after, After Hurricane Maria too. And I, I would like to think. I don't know that it's really the case, but I would like to think that. That we have a certain set of kind of decolonial experiences that that might be more effective than. Than what might be the equivalent resistance in, say, the University of Massachusetts or, you know, the University of Texas, because of what is already established here.
- [38:53]
Emily Bender
Yeah, I love that question. What have you been reading? Is. Is beautiful. And I think one question is how do we. How do we get there without a natural disaster? Because, I mean, I think it's wonderful that that is part of the reaction to surviving Maria and everything that came after that. But as you were saying, I think that the. There's a lot that Puerto Rico can teach the rest of the world and the rest of the US and what have you been reading? Sounds like a great start.
- [39:21]
Alex Hannah
Yeah, that's, that's great. I also worry that, I mean, that's also one thing I do worry about with LLMs is that the kind of, the kind of reading. And I think we were already saying that without LLMs. I remember teaching a few years ago, it was a little surprising that. Yeah, I think, I think the crunch on student time and the kind of discouragement. Discouragement of reading more generally. So, I mean, it's, it's. It's. It's a really difficult thing too, especially if this is something that students are tasked with or there is kind of. There's alternative demands on attention. So. Yeah, I mean, I don't have any good advice on how to cultivate the reading culture other than try to do with other people, because I think that's the best way to do anything.
- [40:17]
Jeffrey Hurley Humera
So.
- [40:18]
Alex Rivera Cartagena
Yeah, I was thinking in some other countries we are seeing that in schools or generally they're limiting the use of social media to the age of 16, I guess. I think. Yeah. So I don't know if it is possible that using some limitating laws, we can say that we can coach people, especially young people, to use more books, etc. But I don't know if it is the perfect method, but it is a trend that I'm seeing. So I don't know what, what you think about that?
- [41:04]
Alex Hannah
I know that. I know on social media there's at least on Tick Tock there's kind of a move for Booktok. But I mean I'm, I'm going to get quickly in over my skis and I'm going to sound very old if I talk about TikTok and so I'm not going to do that. Instead I think that. Yeah, but I do think there's a way to encourage folks to do that. And it's. I think there's an interesting opportunity too that is raised by the use of LLMs and the use of chatbots where I think there is a bit of a cultural resistance where I think, you know, there's, it's almost gone the other way where some young people or some people think like, well, I'm actually going to go and try to do this more slowly and I will read as a bit of a resistance element to it. Of course, you know, that's. I think there's at least enough culture, enough of a sort of uncoolness in this. Again, I can't determine what's uncool. I'm dramatically uncool. But it is certainly a way and that I think there is a boomerang effect of that. I hope.
- [42:15]
Emily Bender
Yeah, I think that's a good hope. But I also think that generally bans are probably not going to be effective. I think that pushing more on supporting connections, supporting authenticity and then in fact requiring accountability of the tech companies rather than banning young people from using this technology. I think if we hold the tech companies accountable for the harms they are doing, that is going to be a more effective direction.
- [42:45]
Alex Rivera Cartagena
Yeah, yeah, I can see that point because like you say, maybe really I'm not that fan like you say that to be banning some technology that I think that can have some good value itself, but maybe a different approach to using this technology will be a better approach to this situation. I think it is a complicated situation. But now I wanted to ask you something that it is more correlated to in Puerto Rico, especially in our campus in Mayaguez. You can situate this argument with Puerto Rico today. What I can read specifically students in Mayaguez today use, learn and apply from your book in their life or academic career.
- [43:58]
Emily Bender
You want to go first, Alex?
- [43:59]
Alex Hannah
Yeah, that's right. I mean, well, I mean this is, yeah, this is a great question and thinking about how to make it very local. I mean the thing that I think y' all have been talking about before is really what existing civic participation that y' all have in your community. And I think it's fantastic to hear how vibrant civil society is in Puerto Rico and Guayaes. So I think one thing that folks can start thinking about is what are ways and asking each other what are ways that there is an imposition of these technologies on your everyday life. Where is it happening? Is it happening at work? Is it happening at school? Is it happening in the doctor's office? Is it happening before you even, you know, apply to a job? So first, identifying where that's happening. And I think that's one place where the book becomes very helpful. We've got a set of six questions in the book in which if you are confronted with a techno the technology or confronted with something that is marketed as AI, what you can ask about it? First, is what's actually being automated? What are the processes that are. Are that the makers say it's going to be streamlined. Second, what are the inputs and the outputs? What do you put into the system and what comes out? And can you connect those things to each other? Like if something says it can look at your face and determine if you're a criminal or you're qualified for the job, that is certainly one, not something that can be discernible from one's face and that should be rejected outright. Third, how is it evaluated? What is guarantees of that it's going to do what it's saying? Who's making that evaluation? Fourth, are the tools being anthropomorphized? Are they being described as human? If they're being described as human, why are they being wired? What's behind that? Is it suggesting that it could be like an AI tutor or teaching assistant? Then what kinds of services in the education, in education and the institution are being devalued? Fifth, what are the processes and labor practices and data practices? Who's working on this to guarantee that the people that have created this technology or if they've had any annotators or data laborers, that they were treated with good working practices and what the training data is of these systems. If those, you know, if those questions or those answers are not available, that's concerning. And lastly, who's accountable? What happens when it goes, when things go wrong? And then who does it serve? And so those are questions that we have in the book that can really be used in many different contexts and situations. The thing that makes us so great in the Puerto Rican context is the vibrant civil society is the democratic society, is the groups that are already formed and those can be really powerful pathways for political education. And for. For really ensuring that there's a integration of society and tech that works for people.
- [47:38]
Emily Bender
Thank you, Alex. I think that covers it really nicely. I want to touch on, Jeffrey, what you were saying before about the linguistic distance that you have there in Puerto Rico, which I think is really valuable. And I think that there's two things going on. One is Spanish is an important language in the mainland, too, but not nearly as much as in Puerto Rico. But also the Spanish, as you were describing when you were in Spain, that gets represented in language technology is probably a little bit distant from Puerto Rican Spanish. It's probably not valorized by the kinds of things that end up on the Internet, the kinds of things that are likely to come out of synthetic text extruding machines. And I think there is a little bit of a trap that it's worth being aware of, which is to say this technology doesn't work as well in languages other than English. I think Spanish is probably near the top of the languages other than English in terms of how well it works, but it's also not going to work as well for Puerto Rican Spanish as it does for Iberian Spanish. And so rather than focusing energy on getting this technology that's actually harmful to work better for the Spanish as spoken in Puerto Rico, I think it's worth more using that distance as an advantage and saying, hey, this, we can take a little bit of a distance from it. It already looks a little bit bad to us. Let's look further into what the possible harms are and what questions we should ask before using it.
- [49:00]
Jeffrey Hurley Humera
That is awesome, Emily. In fact, I feel like you just read my mind at what I was, What I was putting down in my notes, what, what I would. What I wanted to. To say and to build on what you just said. The way that English exists on our campus, I mean, people speak it. It's not. It's not banned or anything. But when it does, when it is used, it's used in certain context and it has a certain. I would say sometimes it comes in almost as like a. To add a. Like a joke to something. It is like a rebellious nature and you play with it. So it's not. Even though it has to be there because of the US Intervention. It is. I feel like there's a very. Like you said, I guess you were quoting me, but like this kind of healthy distance from what the imperial model desires. And that's. That has been developed over time. And I feel like if we were to think about technology in the same way, that's. That would be a. That's I think why. I think it will be a really interesting future conversation that we're going to have with Alex and hopefully you guys will come to Puerto Rico. We'd love to have you. And to think about, you know, these technologies in a specific place like ours because it's very unique with not only the, the monolingual Spanish, but also with Spanglish because it's. Spanglish is also a, a very important part of, of, you know, the way that we communicate with it in. On the island. And so thank you guys so much. I know we're. We're just about out of time and. And this is really. This hasn't wrapped up our conversation. This is. I feel like it's beginning of. Of. Of a larger one. And thank you guys so much for, for these discussions about racism, classism, eugenic policies and how these shape the histories, but also raise these consciousness about these in AI and it's, it's, you know, we really need to, to step back and to, to really think critically about what's happening and who's benefiting. And I feel like your guys book was really wonderful introduction to that and, and it's the beginning of a, a larger and much broader conversation that Alex and I are going to continue soon on. On the. On the second episode in Spanish. So thank you guys so much for, for today.
- [50:45]
Emily Bender
Thank you so much.
- [50:47]
Alex Hannah
Thank you. It was such a pleasure.
- [50:49]
Alex Rivera Cartagena
Yeah, thank.
- [51:02]
Emily Bender
You.