Jazz in the Age of AI

Michael Ambrosino hosts this episode of The Buzz, a discussion with Josh Antonuccio, Matt Powers, and Jon Irabagon as they explore the effects of AI on journalism, music, production, education, and jazz. 

Josh is an associate professor within the media production and recording industry major in the School of Media Arts and Studies at Ohio University. He's worked extensively within higher education since 2007, establishing innovative music and media industry curriculums and developing the expansion of experience-based music industry education. He is also the director of Ohio University's Music Industry Summit.

Jon is a multi-reed instrumentalist, composer, arranger, band leader, and faculty member at the University of Illinois in Chicago, where he teaches jazz saxophone and courses in jazz history. Winner of the 2008 Thelonious Monk Saxophone Competition, a Rising Star Award in DownBeat magazine's alto and tenor saxophone categories, and a recipient of the Philippine Presidential Award, Jon's latest album is "Server Farm," a musical exploration of how artificial intelligence affects our lives.

Matt is a professor at the University of Washington's Department of Communication, where he's the co-director of the Department's Center for Journalism, Media and Democracy. His book, "The Journalist Predicament: Difficult Choices in a Declining Profession," explores journalism within the transformations confronting the profession. He's also the co-editor of "Rethinking Media Research for Changing Societies," exploring how researchers can make sense of the massive changes confronting politics and the media.

Take a look at the Jazz Omnibus, the 600-page anthology of 21st century photos and writings by members of the Jazz Journalist Association, available online and wherever books are sold

Don’t miss new episodes of The Buzz. Make sure you follow us wherever you listen to podcasts.

For more from the Jazz Journalists Association, go to JJANews.org.

Transcript

Michael Ambrosino: Hello and welcome to The Buzz, the podcast of the Jazz Journalist Association, an international professional organization of writers, photographers, and broadcasters focused on jazz. I'm Michael Ambrosino, jazz journalist, broadcaster, audio producer, and founder of 33third.org. I'll be your host for a discussion with Josh Antonuccio, Matt Powers, and Jon Irabagon.

As we explore the effects of AI on journalism, music, production, education, and jazz, Josh is an associate professor within the media production and recording industry major in the School of Media Arts and Studies at Ohio University. He's worked extensively within higher education since 2007, establishing innovative music and media industry curriculums and developing the expansion of experience-based music industry education. He is also the director of Ohio University's Music Industry Summit.

Jon is a multi-reed instrumentalist, composer, arranger, band leader, and faculty member at the University of Illinois in Chicago, where he teaches jazz saxophone and courses in jazz history. Winner of the 2008 Thelonious Monk Saxophone Competition, a Rising Star Award in DownBeat magazine's alto and tenor saxophone categories, and a recipient of the Philippine Presidential Award, Jon's latest album is "Server Farm," a musical exploration of how artificial intelligence affects our lives.

Matt is a professor at the University of Washington's Department of Communication, where he's the co-director of the Department's Center for Journalism, Media and Democracy. His book, "The Journalist Predicament: Difficult Choices in a Declining Profession," explores journalism within the transformations confronting the profession. He's also the co-editor of "Rethinking Media Research for Changing Societies," exploring how researchers can make sense of the massive changes confronting politics and the media.

Matt Powers: Good to be here.

Michael Ambrosino: I want to set the table for our audience because obviously this is a very large topic, and AI is pretty much embedded into anything and everything. Specifically, we're going to try to focus on how AI has radically changed how we look at journalism, media, music, production, and education as it relates to these fields, as well as take some time to explore from Jon about the role that AI played within the context of creating his new album "Server Farm."

So as we start with journalism and media, Matt, for years now, the news and media landscape has been radically altered by our digital economy. Is it fair to say that the use and inclusion of AI has only accelerated that trend? And if so, what are some of the major factors that you've witnessed over the past five years?

Matt Powers: I think artificial intelligence very broadly is contributing to changes that have happened in journalism. But as you note, Michael, these are changes that have been going on for a long time. The biggest issue is just the changes in the business model and the political economy of journalism, which has made it increasingly difficult for news organizations of varying scales to actually make money and to be sustainable. The biggest issue for news organizations and for journalism very broadly is simply the changes in the business model.

Michael Ambrosino: Josh, what overall trends do you see in arts and cultural journalism over this time period?

Josh Antonuccio: I think we're starting to see the first fruits of it. To Matt's point, it is being radically altered, I think in large part because of the business model changing. I think these partnerships that you're starting to see right now with Condé Nast, The New York Times, and others partnering with AI companies, or at least leveraging their data and their content, is a big harbinger of where things are going because these AI systems have to be trained on something. And those that control that content—that being the media companies—have a huge card to play by creating these deals that are now in motion over the last year or so. We're going to see more and more of that to come.

Michael Ambrosino: I'm curious because I remember the first time I saw one of my favorite drummers, Eric Harland, basically go to a software company so that they could sample his entire sound, his brand. It made me think, why would you give that kind of agency away? Do you feel like media companies are in some ways selling their soul as they provide proprietary content to train an AI model that could replace their reporting?

Josh Antonuccio: I don't know if I'd call it selling their soul. I think they're being put into a position of survival, and that's why on the music side right now, you see the three major labels suing Suno and Udio. But I think they are setting up really as what happened in the past with folks like Napster, where they're going to have to strike some kind of deal because this technology—right now we are at the point where this is the worst this technology will ever be. And we've only had it in a consumer-facing, accessible platform—that being ChatGPT—for less than three years. So it's only going to expand more and more. And so I think it's totally a survival move.

Michael Ambrosino: So Jon, you and I both know this. I'm a jazz journalist, you're a musician, and jazz really depends on cultural reporting, as arts and cultural reporting influences dimensions to some degree. What effects do you see this having in terms of the music that you love, and in your view, has any alternative reporting taken its place?

Jon Irabagon: For my friends and I and all of us that are touring around and trying to make a living playing creative music, it's this huge juggernaut and this huge tsunami wave coming for us. And the only thing that we can do is really sound the alarms, especially as musicians, because we're seeing Spotify takeover, we're seeing AI music being created and celebrated on Spotify as opposed to the real thing. A lot of friends of mine just have their heads in the sand and they're like, "There's nothing we can do individually." But this is our chance to say something because it's only going to get worse. And it's interesting and cool for me to be here with you all because the AI thing—it's coming for everybody in all sectors of life. But music and journalism are among the two first professions or artistic endeavors that AI can more easily replace.

Michael Ambrosino: It's so funny you say that because, Matt, I was reading through a preview of your book and I was struck by how the conditions you describe—declining work environments, stagnant pay, reducing cultural agency and authority—it reminded me so much of what I've seen jazz musicians experiencing for years now. Conversely, has any of this disruption that you've studied and reported on helped advance or improve journalism in ways that were surprising to you?

Matt Powers: Yes. I think it's true that every crisis is also an opportunity. The question is always an opportunity for who and under what conditions and for what types of work. Within journalism, of course, there are people who are doing interesting things with artificial intelligence of various sorts and finding ways to actually differentiate themselves and their work. One of the risks of these types of technologies, as both Josh and Jon have pointed out, is that there is a kind of de-differentiation that happens because, of course, all of this is based on patterns and it's based on being able to predict things. And you base predictions on how things have worked in the past, whether we're talking about insights that a cultural journalist might have about a jazz album that they really liked, or an investigative reporter that's able to spot a pattern that they hadn't seen before that actually requires seeing something different that people haven't seen. And so I think oftentimes the people who are able to use these devices and tools in interesting and novel ways are doing it precisely because they're cutting against the grain.

Michael Ambrosino: So Josh, journalism is seeing radical shifts in its structural foundation and how it operates. I remember when I was going to get a master's degree in journalism that all of a sudden I saw things like data forensics and computational journalism. They're just two of the new disciplines that come to mind where computer sciences morphed into things that journalists need. As director of Ohio University's School of Media Arts, what are some of the ways that you're trying to reimagine what journalism and communications can look like moving forward?

Josh Antonuccio: We run tangential to journalism. There's a separate school for that in our college, but insofar as we interact on the media side, I think all of us are thinking very much about how to train the next generation of creative professionals with tactile and very applicable AI skills. Ohio State just announced that they are going to have every student at their university learn AI in the coming semester. And if you see some of the most recent warnings from folks like Anthropic and others that are talking about the next wave of job losses that may be on the table, the percentages are staggering what they're talking about in terms of these kinds of entry-level positions that students used to take that now AI will very likely be able to start doing. And again, you need somebody to know how to use AI. I think this combination of AI—artificial intelligence—and HI—human intelligence—is where we're looking to really meet the need in the moment, which is we have to be able to train students for the world they're entering into, not the one that we came from. And so that's what we're really aiming to do.

Michael Ambrosino: And just to give context for what you're talking about, the latest report from McKinsey that I think came out in late 2024 said that right now it's slated that AI could take over 300 million jobs worldwide. So it seems like an appropriate way to incorporate AI into your pedagogy as opposed to not give students that opportunity to compete in the marketplace.

Certainly, it feels like everyone now has a mic and a video camera and a variety of engagement strategies to take content and monetize it. And this trend has played a role in displacing traditional journalism, even as it paves the way to give someone who wants to take journalism seriously a way to monetize their effort. I'm curious, and you can all jump in on this, how do you all view the growing power of various influencers and the changing nature of how society engages with their work? Matt, you want to start us off?

Matt Powers: Of course, the category of influencer is huge and it's plastic—it shifts quite a lot. If I had to make a generalization about it, I would say that the people who tend to be identified as influencers are generally people who find a way to adapt their skills to whatever a particular audience wants to see. And so oftentimes it's quite entertaining rather than super serious. It's a way of giving people the information that they want, when they want, how they want. And so it's very much driven by a sense of being a sort of service person in people's everyday lives, which in a big picture way is a very different way of thinking about cultural production and journalism in the past, which in some sense—to overstate the case a little bit—was oftentimes more like "Here's what a teacher thinks you should know." And so I think we have this shift now that's very audience-centric, that has many wonderful aspects, but that also raises these questions about what's getting left out.

Michael Ambrosino: Jon, if I give you a specific context—and you and I have talked about how traditional jazz magazines, for example, can feel like walled gardens. They're not necessarily going to deal with the most progressive aspects of the music, or some of the more radical storytelling within an album. So Jon, have you seen online people who are beginning their own traditions that kind of saddle up close to journalism that are actually filling that void?

Jon Irabagon: On the performing side of things, lots of musicians are not into doing social media and the influencer thing at all. And lots of those musicians are getting left behind because the venues who are getting squeezed, just like every other aspect of the middle class in today's society, they're wanting a musician that's going to come in and guarantee ticket sales. One of the last things on their mind is, "Is their music good?" The venues are not thinking anything about that. They're wanting to get butts in the seats. And so I've seen many instances where an influencer who has 50,000 followers or something, they're going to get the festival gig or even just a club gig over someone who's been out hitting the pavement for the last 20 years, strictly because of the amount of social media followers they have. So it's definitely an uphill battle.

Michael Ambrosino: Josh, have you seen how AI has empowered independent journalism as a thoughtful alternative to traditional news outlets?

Josh Antonuccio: From my vantage point, I would say that social media is the way that discovery is really fueling this next generation to discover jazz. Traditional jazz magazines or any of the traditional outlets—they're going to be on outlets such as TikTok looking to discover. And I think that's what a lot of this is about—how do you connect the great artists, the great musicians of the moment with a new audience or an emerging audience? And I think you have to go to those outlets and platforms where that type of discovery is happening.

Now, I would say that where I see AI assisting in that is certainly in helping to develop strategy in terms of how you engage. Using, I mentioned, Anthropic, Claude AI, or ChatGPT—it's really amazing how much you can program these AI systems to be essentially a social media strategist with you, help you to develop a marketing campaign, help you to think about where to engage and how to engage. And yes, the art matters, but art is separate from commerce. And when you look to interlink the two, both have to be considered. And certainly if you're not going to make money, those types of opportunities are not going to open.

Michael Ambrosino: So as we move on to education, it seems impossible now not to ask students to use AI within their coursework. Have you devised any guidelines regarding the ethical practice of doing so at the University of Washington?

Matt Powers: So the university does have guidelines, as I think most universities have. They come in different flavors where it's "If you want to allow for AI usage in your classroom, here's some suggested language that you could use. If you want to ban it, here's some suggested reasons that you could use." And the guidelines often come from the top in the sense that it's administrators who are crafting it, and then you can adapt it as you like or see fit for your own coursework and curricula. I think that's the same as what's happening in news organizations where it's mostly coming from management and the journalists are forced to work with it. But I think the key question on a very practical level for me when I think about it is, how do you simultaneously take seriously the changes that AI really affords, creates, demands, while also inculcating in students a sense of skepticism about some of the hype and the overblown claims regarding what AI is and what AI can do? And it's hard to keep both of those things in mind at the same time.

Michael Ambrosino: To follow up on that question, how do you deal with the polarity of plagiarism but also verifying that what an AI agent actually gives you in terms of information is accurate?

Matt Powers: It's a great question. So one of the questions then becomes, how do you actually get students to verify facts, to do things that are actually—should be—very old hat for journalists? And one of the things that we always point out is that actually when it comes to very basic information, a lot of these tools are just flat wrong. And we'll show them the studies that say basic things like if you want to know where to vote in a particular election, there have been studies now done for several years that show that very basic questions in terms of "Where do I go? What date do I go? Am I eligible?" And so then the question becomes trying to teach students about where they can actually go to find the information. So in certain ways, it's not so different from what we said 20 years ago when people started using Wikipedia. "Hey, it's a great starting point, but it's not the end point."

Michael Ambrosino: Josh, I'm really curious—are you seeing ways in which AI has accelerated this gap between being able to research in a fundamentally poignant way?

Josh Antonuccio: A hundred percent. I agree with Matt. I think in some parts it's changing, but it's also accelerating in a way that we don't really have a precedent for. You think about where the internet was in the Netscape era. That's essentially where we are with AI right now. And the accuracy of information, I'd say, is the biggest hurdle at the moment, but the ability to generate wholesale in your own voice, to train a language model to write like you, to get—let's even say 80 percent of your writing—just within your grasp in less than 30 seconds is just a step that has seemed unfathomable even a few years ago.

And furthermore, right now, again, going back to some of the reporting I was mentioning earlier, most people you read that talk about where AI tools are going in terms of replacing jobs—that is starting to happen. But I think it's going to be much more the case, and is much more the case, that it's people that know how to use AI tools and how to leverage them effectively will have jobs over those that don't. And I think the speed at which you can get to something is going to become the new determining factor of employability and efficiency in the workforce.

Michael Ambrosino: Josh, do you see that AI prompt methodologies are replacing more vigorous ways that students approach researching a topic?

Josh Antonuccio: Not necessarily, but I do think that prompt engineering will become one of the most vital skill sets in the near future.

Michael Ambrosino: Jon, you teach saxophone, composition, and arranging. Have you ever experienced an example of where a student used AI within the context of these traditional assignments?

Jon Irabagon: I teach a 500-person general education jazz history class, and the two papers in the class—one of them is they have to review a Blue Note record. Like one of the ways that I'm trying to avoid the AI usage is that I'm like, "Okay, you have to give me a bio on Lee Morgan for 'The Sidewinder,' but you have to review the songs, each song individually with no AI language. And I just want to hear from you. I want to hear from the student individual. And I don't care if there are run-on sentences or just fragments. I just want to hear about what you think about these songs." And the students, a lot of them can't do that. There are already just a few semesters into the ChatGPT era—they're already completely dependent on having the AI tell them what they thought about "The Sidewinder." Each semester, I'm trying different ways to try to get them off of it, but I think what Josh and Matt are saying is that AI's here to stay, and the correct way to move forward is to find a way to get these students to—okay, if they're going to use AI, then find a way to interface with it personally and make those prompts work for them more effectively.

Michael Ambrosino: When you think of the tactile way in which you teach playing a saxophone or composing or arranging, even with examples that you could call from the internet, can you envision AI being a thoughtful or provocative aid for your students?

Jon Irabagon: AI's got very powerful tools, and I think that used in the right context, in the correct way, and with the students respecting the history of how they've been learning and the history of the music, I think there's a possibility to have AI help at this point. As far as teaching a one-on-one saxophone lesson, I don't think it's quite there yet, but I can see a future where it could help. But I don't think, especially in jazz, which is such a creative and personal art built on personal expression, I don't think it's ever going to fully replace a mentorship kind of situation.

Michael Ambrosino: Shifting to music production, Josh, I know firsthand that many of the design and audio tools that I use have AI embedded in them. For example, the way that AI can magically repair bad audio or damaged photographs. Are you seeing AI advance at all into the way that audio and music production routines, including mixing and mastering, are being utilized right now, and has that become part of your curriculum?

Josh Antonuccio: Have I seen advances? A hundred percent. The tool sets that are available now—just taking like text-to-music, although I don't use that as a production tool, but just the ability—I mentioned Suno and Udio earlier—that alone is a massive amount of power. But even ideation tools like Lemonade, BandLab, and others where, again, this kind of acceleration trend where an artist might have had to have really wrestled with something, it just enables people to just get somewhere faster. Now, whether that destination's desirable is another conversation, but just the ability to do that is huge.

And as you mentioned, I see it more on the restoration side or the ability to separate stems, which essentially are like—the best example is The Beatles track "Now and Then," where because of the way they recorded, everything was essentially baked together—drums, vocals, guitar, bass. You can't separate it out of a two-track mix. But AI tools like AudioShake enable you to then separate them, and then you can address them differently or you can license them out for beats or do other things with them. And that is changing, and I think that is going to become way more the norm in terms of responsive AI tools in mixing and mastering. Again, that will get you somewhere much faster. Say I just need a preset, something that would get your EQ to EQ for a snare or something. Imagine instead: "I want this snare to sound like Nirvana's 'In Utero'" like right now, and it will just give you that. Technology is not consumer-facing yet, but we are very close.

Michael Ambrosino: Are you at all trepidatious that we're offloading hands-on knowledge to an AI system that definitely needs to be learned by a person who's trying to become a studio professional?

Josh Antonuccio: Not really. I think maybe I might have been for a minute, but I talked to some of the people that have created some of the most iconic records, and they just haven't shared that trepidation. And also, if you walk back through recording technology, it is the same thing people were saying about Auto-Tune. Same thing people were saying about Pro Tools. "That's not real recording because you're not physically recording it to something like a tape machine," or before that, direct to vinyl. I mean, there are concerns I have, but technology is technology, and there have been complaints about it all through time. And I think that it's just going to present new challenges and new opportunities in ways that we could have never fathomed even in the last few years.

Michael Ambrosino: The question here I have for all of you, because you're all creative in one facility or another, can you name three things that AI has done to benefit the way you work, teach, or support your own creative activities?

Matt Powers: Well, it certainly can make some things more efficient. And so I think like anybody else who teaches classes and writes, you can get stuck in your head. And sometimes if you say, "Hey, I'm working on this, can you show logical inconsistencies?" That's great. And I like that. Also, if you prompt something like ChatGPT to say, "Hey, I'm wondering what this particular person would say in response to what I'm writing here," you can begin to envision dialogues that aren't actually happening. All of that I think is actually quite fun and quite interesting. I think the dominant tendency with respect to AI in general is a very utilitarian approach: "What's the least I have to do in order to get the most gain?" And I think that's a recipe for boringness just in general, whether it's student papers, whether it's cultural production, etc. And I think probably the more interesting uses are going to be ones that have to do with, "How can I try and do something better, more interesting, having more depth than I would do if I weren't using this tool?"

Michael Ambrosino: Jon, what about you? I mean, you've created a whole album on the subject matter. Has this made you embrace more AI tools in the process of being a creative professional?

Jon Irabagon: Going along with what Matt said, I think there are a lot of possibilities there. There's a lot of power there, and if I were in more situations where I could utilize it, I feel like I would because of all the possibilities there. For me, just being more or less a traditional musician that's trying to improvise and connect with the past and find my own voice within it, I just haven't had to dive in that way. But with this new record, I did get into AI way more than I would have, and I tried to recreate a way that I could be an AI just to see what that would even mean and entail for composing for this record. I think down the line for the next edition of the "Server Farm" series, I'm going to try to incorporate it in other ways just to challenge myself and try to see what's going on.

Michael Ambrosino: Josh, I can imagine you use a lot of AI tools to streamline productivity within research or whatever. In terms of being creative yourself, have you been surprised by some of the elegant ways that AI has helped you be more creative or enhanced your creativity?

Josh Antonuccio: I just put a record out last month, and I found that AI helped me to be able to take more of the mundane tasks off the table so I could be more focused on the creative. I found myself using tools—I'm a big Midjourney fan, so some of the storyboarding for the short-form videos that we did was really helpful. But I wasn't using my AI tools to help me write. It was more like, "I've got these other tasks I need to hire, or I'd like to have assistance for," and that's the way I've thought about it. Greg Brockman—he used to be at OpenAI. I saw him at South by Southwest, and he really helped change my thinking about this. If I had the budget, I would hire assistants to do all these tasks, but I have AI assistants that can help me so I can be more focused on the creative process, my own writing, my own recording, and so forth.

Michael Ambrosino: And that is a methodology that you see throughout the branding of AI—that you can have AI agents, especially if you're a small business, that go out and can book your tour, can help you write the promotional one-sheet, can do the things that you simply can't afford to hire out to get done. So sticking with the idea of art and the process of being creative, Jon, can you walk us through why you decided to create a project that focuses on the nature of AI? And what kinds of research did you do in preparation to record your album "Server Farm"?

Jon Irabagon: For me, this was going to be the first record I'd ever written for 10 musicians, and I wanted to find a way to not overwrite but also write specifically for these other nine musicians, plus myself. These are nine of my favorite musicians, and they're great free improvisers as well as chord-change players and everything. So I wanted to find a way to allow them to be free from the pieces of paper as much as possible, but still have five compositions of my own address this oncoming onslaught of AI in our everyday lives and incorporate some version of that in a more organic way to myself.

So I was like, "What would AI do if I just told ChatGPT to compose five pieces that involve these 10 musicians? What might it do?" And I realized, "Okay, it might go through the entire history of all 10 of us and find all the records we're on and then come up with—spit out—these compositions that incorporate different bits and pieces of all of those records." I found AI could probably do that in what, a minute or 10 minutes or something like that. But I took 10 months, and I found every record that Wendy Eisenberg's on, and I found every record that Miles Okazaki is on, and Peter Evans, and so on and so forth. And I went through a great 10-month odyssey of listening to all these records, taking notes. I didn't transcribe anything like, "Okay, he plays this lick over and over again, so I'm going to use that." But I took notes on the tendencies, maybe certain gestures or certain atmospheres that they liked to improvise in. And I took notes of all of those things for all 10 of us, and then the compositions—I went back in a very traditional, old-school way, just composed myself. But with these notes and that 10 months of intense listening in the background, and the pieces came out very much new versions of pieces that I've written, but with these new nuances that I haven't traditionally used in my own compositions because it's been influenced by this 10 months of listening.

Michael Ambrosino: Did you have any similar process with "Server Farm"?

Jon Irabagon: I was subbing in Maria Schneider's orchestra several times and for a couple tours during that period. So I was in the rehearsal room with her, I don't know, maybe a dozen times while these pieces were being developed. So I saw her speaking with Clarence Penn, asking for specific things on the drums from the snare to emulate binary code or whatever. I wanted the nine musicians to be as pure and untethered to anything as possible, but in passing conversations, they knew that it was going to be about this AI thing. And so each of them brought their own histories and their own research to it. But I didn't want their side of improvising to be weighed down by this topic.

Michael Ambrosino: Josh, Wired magazine did a recent piece on two musicians, Mike Smith and Jonathan Hay, who ironically created an album called "Jazz," and it spoke to the fact of deep-faking a billion streams. And a while back, there was also an example of a fake Immanuel Wilkins album—who is a young, superb alto-tenor saxophonist—on Spotify that was up on Spotify for two days before it was taken down. So I'm curious, can you give us your take, and especially if you have any worries about the effects of hacking digital systems that are used to determine success in music or even the role that synthetic music might have on the economy that supports the music industry?

Josh Antonuccio: Yeah, I and everyone else have huge concerns about that. Most estimates you read, it's somewhere between 130, 150,000 tracks a day uploaded by actual musicians to platforms, and that number just exponentially inflates when you incorporate AI-generated music, those deep fakes that are being uploaded that were mentioned in that article. Shout out to Liz Pelly, who's also done great work on this subject. She wrote a recent book and had a great piece just recently. But all of this is to say that it is very hard for musicians to negotiate that in their own economy of understanding how to get discovered and how to stand out amongst an amount of noise that wasn't even in the imagination of people just even five years ago. And it's not in Spotify or other platforms' best interest necessarily to stand up for musicians—they are businesses.

Deep fakes are one issue that they are concerned about, more so that if they're able to generate their own mood music or something like lo-fi beats, which is such a massively popular genre, or just people that just want something in the background—or you look up a coffee jazz playlist—if they control the rights to that, then they're collecting off of that. But it's at the detriment of other young, emerging, or other jazz musicians or other musicians that are looking to get discovered or to get some kind of compensation, which is already very hard to do, to generate the amount of numbers you need to make money. Nearly just about a quarter don't even get listened to at all. So just the point of discovery alone is hard, but with this much extra content, it's very difficult.

Michael Ambrosino: Amanda Ginsburg, who's a writer, photographer, and editor at The New Yorker, she wrote a fascinating piece on Substack that outlined an interchange she had with ChatGPT asking for editorial guidance. And it turned out that the AI agent failed to read her pieces thoroughly and then repeatedly lied about it. Matt, when journalists work with AI, how can they verify the authenticity of any information or editorial guidance that they receive?

Matt Powers: I thought that exchange was fascinating and great at pointing out something that AI is uniquely not good at doing right now, at least within the world of text and writing, which is actually offering interpretations and analysis of things that humans actually do. It's not good at actually being able to go through someone's writing. It can mimic the sounds, it can mimic the texture. But it can't mimic the complexity of thought because it's not in and of itself thinking. I think it's a category error to think that editorial judgment is something you could get from AI at this point in time. And so I thought it was a fascinating exercise, but one in which it would make a lot more sense to actually use AI for different reasons, which is to say, "Hey, here's what I'm trying to say. Could you suggest three ways in which I could shorten it or make this punchier, or different headlines," etc. All things that I think are reasonably good, but it would not be my first inclination, second inclination, or third inclination to outsource ethical and/or editorial judgment issues to AI.

Michael Ambrosino: Josh, I'm curious—the music industry summit that you're a major part of took place last March. What was the vibe? I noticed that there are a lot of young musicians, there's a cross-range of different kinds of genres. What was the vibe in terms of the way that AI is affecting the music industry, and did people have any profound observations on how AI could reshape the industry overall?

Josh Antonuccio: Yeah, AI's been a part of our conversation at the summit. This was our seventh year, so for the last three years effectively, and we've had AI music partners like BandLab with us to talk about this, to do demonstrations. We've had folks from LANDR, which is an AI mastering tool. It's been around for a minute. I think in general there is concern, but there is also—young artists want to make art. They're dealing with a unique opportunity and challenge set that's certainly unique, but still they want to figure out how to make it work, and they are hungry. And many of them are just interested in how to use these tools to advance that ambition. And so I wouldn't say there's a lot of concern, but everybody's just trying to figure it out. There isn't a lot of doom and gloom, but there certainly is a lot of attentiveness to the moment for sure.

Michael Ambrosino: I worked at an arts and educational institution for two decades, seeing wave after wave of artistic processes be augmented and sometimes completely consumed by digital processes. And the fear and promise that brings to the variety of artists that support every sector of the creative economy. One thing I don't understand is why humanity has such a desperate passion to explore ways that technology can fundamentally augment or even replicate human creativity. How do you all view this phenomenon? How about Jon? Let's stick you with a hard question.

Jon Irabagon: I think every generation—like we've said, every generation's trying to deal with what the tools are that they've got in front of them. Joe Zawinul was taking brand-new synthesizers on the road, and he was lucky enough to have a great roadie that knew how to fix them and everything, and he created some of the greatest music of all time with Weather Report. But there haven't been that many other Weather Reports. So it's going to take—on the performance side of things, it's going to take complete genius. There'll be one or two every generation that can take this kind of information and make something completely artistic that still is reasonably humanized. And when that happens, it's going to be amazing. And it's going to be a complete ground shift for everybody.

Matt Powers: I guess I'm not so concerned that human creativity is going to be stamped out. I guess I have questions about the social organization of ways to ensure that human creativity flourishes, that it can actually provide things like careers for people who are willing to go into it. No one's ever gone into cultural production because they seriously thought they were going to get rich. This has never really been the case. But it was the belief that if you worked hard, maybe if you got a side gig doing something else, that you'd be able to have some kind of workable middle-class existence. And this is across all cultural occupations becoming less and less the case. It's not that I think human creativity is going to go away—it's that I don't think societies do a very good job of supporting the types of creativity that is actually essential, not really for being efficient, not really for any of that stuff, but just for, in the most basic way possible, having a life that's interesting, fulfilling, and meaningful.

Josh Antonuccio: Yeah, I would totally agree with that—that I think it is very difficult to exist as a middle-class musician right now, purely as a middle-class musician. Alongside of that, I think there are a myriad of opportunities for those that want to do art alongside of other things and have those tool sets to be able to do that. But I think that is the biggest challenge right now based on just a number of economic factors that—some we talked about—streaming, the live music industry coming out of COVID and what that has meant. Where the middle has gotten squeezed out and venues have struggled and some have come back, and the independent organizations like NIVA have done a great job supporting independent venues. I'm not concerned—I would agree—I don't have a concern that creativity, human creativity, pure human creativity is going away. But I do think we are entering an era where we will see hyper-generative output like we have never seen nor been able to experience before. I think some of that will be new, but I think a lot of things are going to be recycled based on previous IP. Just a quick example—you look at what the BBC just did with Agatha Christie, literally using an AI system to have her teach a class on writing, all fed with her original words and from her. But that's just the beginning of, I think, where we're going, where people are going to begin leveraging IP that is used by an AI system in some output that people want to engage with. Again, just top of mind—would you want to take a sax lesson with John Coltrane? Now, there are all kinds of ethical issues with that, but if the people that own his content, his likeness, want to license that out to teach a new generation about jazz, you can see very easily where this could take off. But again, it becomes at the expense of new and emerging middle-class musicians that are trying to come up. It doesn't let people get out of the way so the new generation can come through.

Michael Ambrosino: Also, there is not a lot of transparency in how these systems operate. Literally, the physics of them, the math, and the data sets that are used are often very reflective of the cultural biases that exist. And when you think about the people who control these incredible services, they are sincerely the same crop of monopolies that we've seen in any other sector. And this is something that I'm very concerned about—it is very hard to tell when we use these tools whether it's empowering my own financial and cultural resilience or somehow helping to erode it. Thank you all so much for joining me and having this insightful conversation. I truly enjoyed it.

Matt Powers: Thanks for having us.

Jon Irabagon: You're welcome.

Josh Antonuccio: Thanks.

Michael Ambrosino: Don't miss any new episodes of The Buzz. Make sure you follow us wherever you listen to podcasts. For more on the Jazz Journalist Association, go to jjnews.org. I'm Michael Ambrosino. Thanks so much for listening.