The AI Journalist’s Dilemma: Who Owns the Future of Storytelling?

Julien Redelsperger: And I'm super happy today to welcome Pete Paschal. He's the founder and CEO of The Media Copilot, a newsletter and podcast about how generative AI is changing the media. Pete has a long career in journalism, previously holding senior roles in global newsrooms such as Coindesk and Mashable. So today we will talk about AI, journalism and the media industry. Thank you for joining metoday. How are you, Pete?

Peter Pachal: I'm great. Thank you, Julien. Thanks for having me.

Julien Redelsperger: Yeah, sure. My pleasure. So Pete you've worn many ads in the media world. You've been a journalist, an editor, a podcaster. What excites you most today about the potential of AI in the media?

Peter Pachal: Yeah, that's a really good question. So AI is bringing a lot of efficiencies. To news production I find, I personally find that very boring because it's it's just about doing more with less. And, that's a mantra that's often repeated in newsrooms these days and a lot of businesses these days, really. But what I really like about AI's potential is if you stack up enough of those efficiencies and you applythem, at a, it doesn't matter, forget the big scale of big media companies, aim it at the solopreneur, or the individual journalist and it's going to enable them and is enabling them to do things they just never could have done before. A journalist, all they want to do usually is tell stories and they'll write a story and they'll publish it somewhere and whether it's on their own sub stack or for the publication they work at.Now AI is the ability to take that and slice it up as many different ways as they want and essentially reformat it, reversion it. Into different formats. So one person now. By telling a story can not just only create a text article, but you can create podcasts and videos and, essentially social media around that story and essentially turns a single person into a massive distribution machine just with the ability to.Take content and turn it into something else. I find that like extremely compelling. And in fact, that's a good chunk of what I do as a solopreneur myself.

Julien Redelsperger: And do you think it is easiest to enter the news industry today as a journalist thanit was, a few years ago?

Peter Pachal: Yeah, I think it is in some ways, and it's much harder in some ways. So it's easier in the sense that you can just start telling stories on whatever platform you want, and then. You can probably much more easily go beyond that, thanks to the power of AI. So if you're a TikToker and you're just doing TikTok stuff and that's all you're doing, you could probably use some AI to take that content and do other things with it in other channels, whether that's a newsletter or your own website or what have you. So in that sense, it's easier. I will say it is much harder now because the media industry has essentially shrunk not just Solely because of AI, but there have been, it's well documented. There's been a lot of layoffs in the space. There's companies due to either down ad markets, bad decisions, or now, with AIaccelerating all this there's just fewer jobs. People coming out of journalism school today, I do feel like. If they want to go the traditional route and go to a publication and cover a beat for a few years and then advance in the latter that's a, I wouldn't say entirely dicey prospect, but there's just fewer opportunities than there used to be 10 years ago.

Julien Redelsperger: What's the relationship between journalists and AI these days? My, my feeling is like it's like a love hate relationship where you can see some leverage, but also you can have questions about the way it's going to change your job. Am I right?

Peter Pachal: You're dead wrong. Yeah. I think it's getting better though. So I think when AI first came on the scene, which is to say when chat GPT came out and showed, you could simply get an AI to write articles for you. I think there was rightfully a lot of skepticism anda lot of people thinking, like worried about it in an existential way, but then also Pick apart what was wrong with the technology.And I think a lot of that's carried over I think for probably a good year and a half after that. There was lots of problematic, shall we say, experiments with AI. So a number of publications tried to use AI to generate articles at scale. It's been shown pretty clearly that's a bad idea. Even though the technology has gotten better. It's a I think that it's had that regardless of the merit of it probably poisoned the reputation of it in a lot of journalists minds. So there's a lot of skepticism about it. There's a lot of weariness around it. But like I say, I think in the last six months, maybe to a year, there's been a bit of a thaw as journalists have. Almost because they have no choice have needed to adopt AI. And I say no choice for a couple of reasons. One, that the tools just have it now. Like it's just everywhere. Every tool, every app with very few exceptions has AI features within it. Now you can question the merit or the value of any specific feature, but they're all there.So you're using AI in some way. And the other thing is that I think journalists have seen that there are really good applications of AI that don't involve it writing articles for you. Like number one research, right? It's just much it's just a, such a powerful research tool. If you go to a search engine like perplexity and get it to summarize research for you. And now there's even more specialized research tools for academic papers and whatever else. So I think it'd be hard pressed to find a journalist that isn't using AI in that regard. And then there'sanother use case, which is at the other end of the pipeline of content. You finished your story, What do you do with it?The past 15 years, what we've done with particularly digital journalists is essentially forced them to be content marketers as well. So it's Oh, you're done your story. Now you got to write your SEO headline. Now you got to write your meta description. Now you got to write all the social copy for all of your networks. And I don't know really many, if any reporters or editors who really like doing that. They justrecognize it as a reality. And and now a good chunk of that work, apart from the final approval of it can be just done by AI. You just hit a button for it. So I think that value add, I think it's indisputable in newsrooms to just use. AI for that sort of chrome around the storytelling practice. Yeah, I think, there's been a bit of a thawing there. It does. Journalists have been convinced from these use cases.

Julien Redelsperger: Okay. So that means it saves time for journalists to do what they like investigating, writing, interviewing people.

Peter Pachal: Yeah. Ideally, like I say ideally because I think in, in, in a well oiled or a well run newsroom where that time for the journalist to tell stories is maximized. Yeah, in theory that, that's exactly what happens. What I think people might assume is that, oh, yeah, now with AI running all this sort of chrome stuff, the journalists like, tell stories and now everybody's happy. Not really because. While all that efficiency is good there's still the reality of the business that How is AI augmenting that in some way? And I think that sort of organizational level, that newsroom and media company level ROI on AI there's a lot of, there's a lot still to be delivered there. And I think a lot of it has to do with creating new experiences. Either on a website or in sort of other channels through AI. And that's we're still at the beginning of that journey.

Julien Redelsperger: Okay. And so for journalists that who are still wrapping the heads around AI, what would be your elevator pitch on how it can actually help them? In their day to day work what do you tell them when talking about AI?

Peter Pachal: There's sort of the carrot and the stick in my view. So the carrot is the stuff I've just been talking about. One, it's going to make your life easier because some of the stuff you don't want to do, you don't have to do anymore. It's also going to enhance your storytelling by being able to help with the research phase. And if you do a slightly more advanced version of that, you can actually tap into the work you're already doing. For example, interviews, research, the stories you've already written, you can start. putting things likethat into a folder and notebook LM or just conversing with an AI by attaching some of that research and just getting better insights at it from the beginning. So that, so in other words it's going to make your life easier. So that's the carrot. And then there's the stick, which is that, Oh, and by the way, if you don't do this, Others will, and you're going to lose your job. So it's basically if you don't adapt and start using AI in these ways and essentially turn yourself into a better journalist, others are going to do it and they already are it's adapt or die.There's always. Someone who wants to hold out and do things the old school way, and I think there's a lot of respect for that with certain aspects of the job, mostly the writing and the storytelling. But, AI is just a reality. It's getting built into all these tools. There's just going to be an expectation that journalists bring AI skills with them to whatever jobs are working in. Yeah, you've got to start implementing this stuff into your workflows and playing with it at the very least. Or the world will change around you

Julien Redelsperger: and based on your experience, how easy or how hard it is to be a journalist today in the U. S. You obviously have a lot to cover but what's the mindset of the journalist these days? What do they think of their situation of the tools of the way they work?

Peter Pachal: Real journalism, I would argue, has always been hard. If you want to be a journalist, And I, power to you. I think that's a great job. Obviously it's a job I've had. I will say the quote unquote real journalism I've done in my career is always the hardest stuff, which is that, okay, you need to really research the hell out of something. You need to know it backwards and forwards. You need to know the important people and you have to call To write a good story like, dozens, if not hundreds of people not just for one story, but like you'recovering a beat, you got to be on it and talking to folks all the time. That's hard work. And hopefully the people around you and the tools you have can get you to take all that knowledge work and turn it into something really great and have, good original information and,even this kind of stuff that someone doesn't want published, which is always like a good definition of journalism. I think that's always been hard and I think in the last 10 or 15 years there's been maybe a misperception or a conflation of content with journalism. So content in some ways is relatively easy. Because if you're just doing content, which is to say stuff to achieve a business goal you can always find yourself shunted into something where you write something for SEO or commerce or a social media network, which was a big thing for five or six years. And you would, you could just crank out like tiny posts that were just. About whatever, I don't know who the latest celebrity is, but it was in when I was at Nashville was like the Kardashians and you can just do that and write the same post everyone else did and enjoy your traffic or you can be a Game of Thrones fan site for while the show was on and get a ton of traffic. Those incentives have all changed. That kind of traffic is no longer reliable spigot and everyone's caught on to the idea that content is not really valuable. So I would say if you want to be a journalist today. It's almost a better time than ever because the actual core of journalism, which is to say unique, everything I've talked about unique information really well researched stuff that's still valuable, if not more valuable than ever, it's just, you're not going to see the same stupid, ridiculous numbers around that kind of content but you have AI now. To not only reversion it as I talked about before, but I think also help audiences find it and reap the benefits of it. That said, we're still working out the economics of how that works. So what I think AI is good at, it's good at identifying unique information and patterns. What hasn't been worked out is how do we make sure the person who's provided that unique information gets the benefits of that disintermediation layer that AI provides? We're starting to see licensing deals. We're starting to see revenue sharing deals between AI companies and media companies to ensure that there is that exchange of value. It's early days.

Julien Redelsperger: And so in your experience, what would be the biggest misconceptions that journalists have about using AI tools?

Peter Pachal: I think the misconception is the similar misconception that a lot of people have when they approach AI, that they think they're getting a finished product and they judge the AI. Based on the output matching what they would have done, which I think is entirely wrongheaded way to think about AI. And I get why it is, there's a almost a personification or anthropomorphization of the tech because it is doing something fairly human, like similar to what a human brain would do. But it's really more like an enhancement. I call my company, the media co pilot, not because I'm trying to fool people into thinking I'm a, I'm an arm of Microsoft, but I call it that because that's the correct approach in my view to look at AI, you're the pilot. It's the co pilot it's taking the cues from you and it's enhancing.Your work. So I think I think the main thing you got to come into is not that expectation that you're going to take the finished product and just use it. It's Oh, it's just getting you further down the road. It's something to bounce ideas off of. And. you need to use it in a disciplined way with yourself as the orchestra the orchestrator the conductor, what's the guy who does the broad as the conductor of thisexperience with this, it's not a being, but it is like a, something that can help you ideate, in the same way a good assistant can. So if you approach it that way, then You're great. It's going to do amazing things for you. And you're going to get, turn, you're going to turn four hour projects into two hour projects. And you're going to turn week long projects and two day projects. It's going to, with a disciplined use of it you're going to be able to be more productive and just tell more stories.

Julien Redelsperger: Okay. Okay. That's interesting because you've been teaching AI for some time now to media professionals. What's the biggest haha moment your students have when they start using these tools, like any good surprises that they discover? Do you have any memories?

Peter Pachal: Yeah, I think it's when you take like early on, it was easy to wow people, right? Cause you would be like about a year ago. It's Oh, wow. It can write things for you. It can do research. It can do. And I don't like, yeah, duh. Now everyone knows that. But today I would say that the aha moment comes. After, I've shown them sort of the basics of prompting and stuff. And then I get into sort of specific tools and automations and it's when they see AI in a in a place where they're already working and there's not this copy and pasting back and forth between chatbots, which is what we were doing with AI, six months ago, we still do it, I still go to chat GPT and pay stuff out for quick things, excuse me, but. When you can integrate it either through an API or through a specific tool and it's just there in Slack, like that's a very super common app that journalists use almost every newsroom I know runs is in Slack. It's all the communication, but if you can, if it's, if you're there and you're brainstorming with your group and there's an AI. They're in the room. Or if you're there and you're bouncing around headlines and the headline tool is right there either in your CMS or in Slack. It's oh, so I don't have to add another step. It's just built into the process. And it's yeah, that, that is really cool. And then you can even take that to another level of automation where it's oh, I, I think it's great that the. AI is going to write social copy for me, but, again even, built into the CMS, you're still got these sort of manual processes around it, but it's what if it, did you literally throw in some zaps and you only maybe even have to look at it, zaps being like one of the automations from Zapier, which I, that's a thing I use all the time. And. Once you see that Oh, it's not just enhancing my work. I can even have it doing things in the background while I'm doing other stuff. So that's like the aha moments, but I think really hit home these days.

Julien Redelsperger: So you're talking about automation, you're talking about API, how tech savvy should journalists be in order to fully embrace the power of AI?

Peter Pachal: I think. Not very, but a decent amount. So I don't really code, even though I covered tech for a long time. And yeah, I learned some basic code back in school. I'm not a coder. I don't do anything in C plus or what have you. I don't even know Python really. And. The thing is, I have been able to build AI apps through these online platforms that are no code. You don't need to know anybody. It helps. It helps to be able to think logically, but you don't really need to do it. There's so many platforms now where you can essentially drag and drop icons pointed at various things, whether it's an RSS feed, that's a common thing or what have you and create some triggers. And all of a sudden you're using AI in these automated ways that you thought you needed an engineer to do before. And you don't even, again, like newsrooms have their own processes, but you have your own processes too. You don't necessarily particularly for the research phase. If you want, say a report on your beat email to yourself every morning. Pretty easy to set that up, in terms of some web scraping and taking stuff from the site and feeds you're probably already looking at and applying some generative stuff to that and boom no codeneeded at all. So I would say you don't like, like the tech savviness is more about having a look at the platforms that are available to you that do things like that within a no code environment, and then just, committing to one. And you'll be able to start building stuff. Honestly, within hours.

Julien Redelsperger: What are the skills or the quality, the expertise journalists should have today? And do you see a change since, when you were working in the media industry?

Peter Pachal: Yeah, I would say for sure. So I think, There was a lot of emphasis put on cranking stuff out quickly and in a short bite sized version during the 2010s, right? That's the content I was alluding to earlier. So I think that sort of, blogging skill set or mindset was,had a lot of currency. I don't think that's necessarily the case anymore. Now that said, journalism is always going to be timely. There's always going to be scoops. There's always going to be some pressure to get your story out as quickly as possible, but I think AI systems, particularly, summarization systems, whether it's AI overviews or perplexity they're looking for something more definitive now.They want unique information and they want something that has cover something from all angles. So I think that inherently takes more work. And so I think it is we used to just call that good journalism, where you just have a story that is a good, balanced, rounded look at a, something that happened or a  topic. So I think that we're probably going to see a return to those roots of journalism type skills where. It'snot necessarily your speed that matters the most. It's your ability to in a relatively quick fashion, give a nice sort of comprehensive look at a story. Will AI at some point enter that picture in terms of helping you with that content? Probably, I think. I think we're starting to see that with brands like ESPN. ESPN is one of the only brands right now that's actually using AI to write articles, but it's not doing it in a cheap articles at scale way that was done about a year ago. They're trying to do it just to fill in some coverage gaps. I see some value in that beyond just. Having more stories. So you have more traffic. What I think as time goes on is going to be valuable for newsrooms is that the way they cover the topics that are important to them is going to become more important. I. P. So if you have what, is would be considered a comprehensive look Over time of an area of expertise that's valuable to either external AI systems or your own AI system. So that's why I feel like this use case of getting AI to write articles probably isn't going away and will probably become more common, but just done in a way that's okay, we're not just doing it to get more traffic. We're doing it to make sure we have good information. that we don't have enough humans to cover about this topic.So you put that in your corpus. And then now you have something like, you see the time AI is a good example, like it was media chat bots, right? So if you have a media chat bot on your site, you want to have information you verified. So as long as you're putting, even if it's AI generated good stories in there that are human vetted and are verified as good pieces of journalism. I think that's going to become probably standard practice in a lot of places.

Julien Redelsperger: Okay. You just mentioned ESPN. That's an interesting story. Based again on your experience and your knowledge, what are the key trends for AI and the media industry? Where do you see AI going into the next year or so?

Peter Pachal: I think video is going to be huge. I think bigger than it is today anyway. So everyone's been predictive. Everyone always predicts video is like the pivot and I'm not saying it's going to be the pivot. I don't think everything's going to take over, but I think what we're going to see is with these new AI. Platforms that the idea of creating videos, video, essentially video versions of articles is going tobecome more common. I think vertical videos certainly have been huge for the past five, six years. I think that's going to essentially come to what more websites. So AI generated videos like some, I think it's AI assisted videos. So That's the distinction. It's a very important one because the thing about videos, you go to a platform like Tik TOK. I think it's a mistake to just look at it and think Oh, people like vertical videos. So I can crank out some AI versions of that and people like it. I don't think that's necessarily true. I think there's an authenticity to like the videos on Tik that you need a human there. But I think what's going to happen is there's going to be seen that this is valuable, but not just. But it's not too expensive anymore, because what you would need to do before is hire a whole team of people and to both figureout your look and your style, your production process and your distribution. Now, I think you need far fewer people. To do all that and with AI assistance for things like background with things like editing. I think you're just going to create a lot. You'll be able to create a lot more video with a lot fewer people. And so the ROI on doing it now makes sense. That's the prediction, right? So you're still gonna get humans doing it. And this will be a big point of contention in newsrooms of people who don't feel like they're talented enough and people who do. And I think some people won't like, not every journalist is going to be doing, vertical videos on their stories, but you're going to see more.And you might see some talented people hired just for that, but again, behind them will, what will be instead of a team of people, it's going to be one editor and with some AI apps.

Julien Redelsperger: Okay. So video that's one big trend. That's for sure. Anything else you may thinkof?

Peter Pachal: I think there's going to be, it's the con The convergence of a lot of the trends I've been talking about, which is that people a greater and greater amount of people will get news summarized to them from, Various places, whether it's perplexity or the site itself, right? Like a time AI will have the that's their new thing. And they're summarizing the person of the year stuff. They're going to spread that to other things. Even if you're not doing a chat experience. You're going to get an AI version of what you're looking at and various, that go to the New York Times or the Washington Post, and you're going to see bullets now and all that's AI generated. So AI is going to become a more of this summarization layer on top of things, exactly what that means at the journalist level. Like I say, I think these engines are going to value more comprehensive content. And I think, there's going to be more of a tendency to write. For to, to optimize summarization.There's going to be, that's going to manifest in big and small ways in terms of how they do things and. Hopefully it doesn't lead to too much overwork and burnout like what I think it's going to do is me. It's going to lead people again to trying to do more comprehensive stories. And I wouldn't necessarily say fewer of them, but I think the density of stories will go down the number of them done by humans, but there will be more valuable over time.

Julien Redelsperger: Okay, how do you think journalists should deal with hallucinations when using AI? So hallucination just to recap, it's just like fake news information dates, people that are created by AI with a very convincing tone. And and that could be misleading, of course. So how do you think journalists should work with hallucinations to avoid it or to double check maybe their work, like what's the impact?

Peter Pachal: Yeah, it's very worrisome hallucinations because one of the problems with them is the automation bias that goes with them because we are all hardwired now as humans to trust automated systems. It's a strange phenomenon, but you see this manifest when people Are told by a GPS that, make a right here, even though like in the real world there's probably not even be a road. It's like the signs will say you can't turn right. And yet they try it was famously portrayed in the office is driving into the lake. So hallucinations, if you see. Sound facts coming out of an AI and they're stating them with a lot of confidence. I think even journalists have a tendency to go the AI said it. So you gotta get past that. You obviously need to fact check anything that you're relying. on for your story. I think that's a given and you should be doing that anyway. This is where journalists, natural skepticism firms serves them very well. That said, here's what I'll say about the whole idea of hallucinations in society now. So we're two years into our relationship with AI. And I think anyone who's even a little bit savvy about AI systems knows they make stuff up. It's been, they just know, it's okay I can't rely on this a hundred percent. It's getting better. I think for most I wouldn't even speculate on what the percentage is, but it's high, like in terms of Oh, it's like in terms of how much it's probably true, like this is the strength of long, large language models. You're never going to completely weed out hallucinations. But almost always it's in the ballpark, like it's not going completely off the rails at whatever your request is. What I, the conclusion I'm comingto with this is that I think readers today and again, talking about things coming out of AI systems are more savvy and they're more forgiving.That they understand that it's okay, like a hundred percent accuracy isn't quite what I'm looking for. And I think they're in a, and I, it's a little depressing to say this, but I think you have to acknowledge reality, which is that I think they're more inclined to still trust what the AI is telling them than the media, because I think there's just been so much trust squandered by the media because of politicalpolarization and all the trends we've seen, they've gotten a lot of stuff wrong. And there's a, some occasionally a lack of humility in that they don't really want to admit it when they're wrong, which I understand. Cause it's okay, then you're giving like fuel to the other side. Butthe whole idea of journalism is you're not supposed to be taking the side. And I think, Because of the incentives of media over the last 10 or 15 years, a lot of places and a lot of journalists lost their way in that regard. There's always been opinion journalism. But it felt like for a while, like everything became opinion. And maybe that's still the case to some extent, but I'm hopeful that, Trends are, a cyclical and thatwe're going to start to see more of a reversion to the mean of what I would consider good journalism, which is just, being fair, but also being honest and presenting all sides of a contentious issue. That's my take on Where hallucinations and trust and how that all that shakes out. I think what that will eventually lead to, particularly the audience relationship with AI that I was talking about a minute ago, is that media companies will be forced to be a little less precious with what they consider journalism, which is to say okay, we're going to slap an experimental label on this. And people will understand that the chat that is coming at them or the information coming at them through this AI layer isn't necessarily 100 percent reliable and that's okay. And I think that last little three, three words, the end, that's okay, has been really difficult for media companies. To get comfortable with, understandably but I think they're going to be forced to because if they don'tget comfortable with it, people are just going to start getting their information from chat GPT perplexity and everyone else. And the media will be here and with their walled gardens being precious about the facts. There'll be a reality outside of that bubble that they won't be playing it if they at least they couldn't do it at scale and they don't use AI to reach those audiences.

Julien Redelsperger: You just highlighted a very interesting story is that we used to get news from, of course, the big newspaper, the big brand like CNN, New York Times, et cetera. Then we moved to social media. And now are you saying that we are moving or we will be moving to get news from chat GPT and and how will that change the relationship between the general public and the journalists?

Peter Pachal: Yeah, I think it's inevitable. We're going to get news from chat GPT. And when I say chat GPT, I just mean, like I say, this AI  layer around things.   So whether you're in some very specific app, like particle particles the hot new. AI aggregator app that uses AI on  hings or Google. Eventually Google, I have no doubt, will reply AI overviews to news, right? It's going to start summarizing news as well. So all of that is definitely going to happen. How that alters the public's relationship with journalists. It's a really good question. I think. You'regoing to have to see this essentially trust and verified sources kind of idea of what is actually going into these AI layers. And that's going to be very contentious. It's already very fragmented because the people who are making deals with open AI aren't the same making deals with pick your service, whether it's meta, Microsoft, Apple, all these sort of. Places where people get news there's going to be a different set. of publishers. Now, I think you're going to see a certain sameness among them. You've already seen publications like Time and NewsCorporation making deals with all, Reuters is a good example because they have they're more facts, like what news stories that are breaking which is a consistent sort of layer that everyone needs. But you're going to see a certain sameness to all these sort of places that are giving AI summaries of news. And it's going to get really hard for smaller publications or independent publications to makeany noise whatsoever in these layers. Like I say, the the economics of this are still being figured out. I'm hopeful that one of these many startups that are trying to build an economy around this starts to emerge as the go to source. And that becomes more of a an easy thing for a I companies to just click a box and include things as opposed to the very cumbersome process of making individual deals with allthese companies because you need to do it at scale, right? Because there's thousands and thousands, if not tens of thousands of publications, maybe hundreds of thousands, if you include all the newsletters, etcetera. And you can't. You can't just sign deals witheveryone. You need to have a marketplace and you need to have it all done automatically. And for example, if you're open AI. And say you want to have comprehensive news about, I don't know the petroleum industry or something in this. And it's of course you're going to have Reuters and you're going to have your, the, whatever major publications that are covering that in your engine. But there's probably a bunch of independent bloggers and, trade publications that are covering that specific industry, and you probably need probably a few dozen of them to really have comprehensive coverage. With one of these marketplaces. They could put their content on the marketplace, self report what it's about and essentially click their box and get, put, decide on the price for their content and open AI can go in there and say okay, like I want to tick the box of petroleum coverage. That includes these, 37 publications, I'm going to, mix and match exactly what I want, but I'm going to click that box, agree to these prices and not these ones. And now, great, boom, now you've got that. So that's where we have to get to. And, there's a bunch of startups. One of them is called toll bit, which has very promising and they're making deals with. Various sort of major publications as well as minor ones that is creating exactly that they just need to grow to a certain size to get the attention of both publishers and the AI companies and then you'll have, then you'll really have something.

Julien Redelsperger: It's funny how the market is changing so rapidly in, in, in this environment, how do you teach AI specifically to news industry experts or professionals that might be on the field for 10, 15, 20 years. Do you see any resistance when when teaching AI?

Peter Pachal: I think that there's resistance when you show people that it's replicating parts of their job that they're quite proud of, essentially creating first drafts or research briefs and like Claude saw in it, it's a, once you tune it a bit and know it give it some examples.It's a pretty good writer. So I think that scares folks. I think, though, once they use AI for a bit, they can know, even though as good as writer it is, there's no there. There's no intent. It's very difficult, I think, for an AI to really put together the facts in front of it and create the insight that is, that humans only really only humans can do at this point. They're getting better at it, but it is, I think there's a there's a fear of this thing that appears to do what you do. And, but once you really put it to work, you realize, Oh, no, I still need to be here orchestrating this thing at the very least. And then You realize that there's a lot of little decisions in your work that lead to the work. And I think AI forces you to realize and evaluate which ones you really need to do and which ones the AI Can do passively and then, there's a, maybe for some of that, a transitional process of getting comfortable with it doing that stuff. Like I say, I think the easy way to convince people, particularlyjournalists, is that, hey, this stuff you didn't like doing, like social posts that's taken care of now. And I think that excites a lot of people.+

Julien Redelsperger: How do you keep track with everything that's going on in the world with AI? We have like new features, new startups, new technologies pretty much every week, if not every day. It could be overwhelming sometimes. How do you keep track of everything to do your job and train, like the people that come to see you?

Peter Pachal: I use AI, of course. So I do some simple things, some of it's just good organization. So I find the best information in terms of what's cool and hot and AI is probably in newsletters. And I subscribe to probably like a million people. two or three dozen of the AI newsletters. Now there's so many of them.

Julien Redelsperger: What are your best three newsletter that you want to share? Peter Pachal: Oh here's the thing. Let me just complete the thought and then I'll answer the question. One is that I have all of those newsletters go to a folder, but I'll put a layer of AI on it so that I get those summarized and so like the good stuff. Gets to the top. So that number one but I really do like the rundown. I like some of the less frequent newsletters like Ethan Moloch's One Useful Thing. Everyone in AI follows Ethan Moloch. He's one of the smartest guys commentating, commenting on AI. His LinkedIn posts alone have inspired a lot of stories for me. And there's a whole bunch of other newsletters. I like the signal cause it doesn't come too often. I also know the guys who do it. But I, yeah I find the most insights. Maybe this is a bias on my part, but from more substack like newsletters that are trying to dive deep on a single topic there's, I don't want to,get them. Oh, no, there's I'll there's the I'll think of another one. That's similar to that. There's one called AI snake oil, which I think is really great, which goes into not necessarily what's wrong with AI, but okay, like they really put certain models and tools through the paces and really come up with some really good insights around it. But I subscribe to like superhuman and all those other ones that are really good sort of roundups and you can get a really good sense of, what's bubbling to the top. in AI from subscribing to a bunch of these newsletters. So that's the main thing. I don't do as much web searching for AI and whatever, honestly, like stuff probably comes to me in a Twitter feed more than probably later than for most I'd like open AI. Whenever they release something, everyone just chats about it, and I'll hear about it whenever. But like I say, my, my newsletter isn't like too newsy. I really do want to dive deep into the stories that actually matter to media and journalists. And so say, OpenAI comes out with a new video model or a new text model. It's what does this mean for storytelling? And what does it mean for media sites? And, how you're going to be able to cover your beat going forward. And if there isn't a clear answer, I just don't cover it. It's probably, probably not it's probably more of a relevance to the tech crowd and developers, which is fine. Those are good people and all that's fine. But it is again, I'm really more just focused on how the practice of journalism and the business of media will evolve. With the advance of AI.

Julien Redelsperger: Okay. Thank you so much, Pete. So at the end of each episode, the guest answers the question posed by the previous guest. Are you ready? As I'll ever be. And after that, you'll have the opportunity to ask a question for the next guest. So here's your questions. Courtesy of Colin Crowell, who is VP North America of Cameleoon, an A B testing and experimentation platform, that platform that uses AI to create better online customer experience. And we can listen to his questions right now.

Collin Crowell: I would like to know what's a misconception in your work that you wish more people understood. I'd love to know how would correcting that misconception change the way they approach the work?

Peter Pachal:A misconception in my work? That's a good question. I feel like one of the sort of the bigger misconceptions That people have is the difference between content and journalism, which I talked about earlier, right? And those are very related, but different jobs and different industries, really. And I do get a lot of people pitching me asking me about, certain opportunities and stories and alluding to can we pay you for coverage, et cetera, essentially. And I know there's sites that operate that way. But the big difference between obviously journalism is you don't do that. That's not to say, you can't have a business model and you can't have sponsors. It's just, there's a you need a clarity of labeling and, you need to make sure that the integrity of your information matters above all. So I think that is a bit of a misconception, a growing misconception because the content industry is ever growing, right? Whereas the new, the journalism one is still human dependent. So it can't really get scaled in the same way with that AI scaling content at large. So yeah that's probably the mostcommon misconception about what I do. It's not really, sometimes people describe me as a former journalist. It's no, I'm still a journalist. I'mjust doing it differently and in newsletter independent form.

Julien Redelsperger: All right. Thank you so much. Peter. Now, what question would you like to pose for the next guest?

Peter Pachal: Yeah, so I. I have a question related to what I just said, actually. So my question is with A B testing and changing how something is presented to a user and gathering the data on that. What is the biggest barrier To having more sites, particularly media sites, adopting that idea. There is I think a certain amount of resistance because people think it might compromise integrity or what have you, but I feel like there should be more of this done in the media industry, but I also feel like it is something that quite isn't in the DNA of just rapid testing. How do you achieve that in places that, are more about journalism and media and what is the biggest barrier.

Julien Redelsperger: All right. Perfect. Pete, it's been an absolute pleasure speaking with you today. Thank you for joining me.

Peter Pachal: My pleasure, Julien. Thanks for the conversation.

This transcription was generated by an artificial intelligence tool. It may not be 100% accurate and could contain errors and approximations.