AI in Journalism as a Co-Pilot

Pete Pachal has a way of cutting through the noise. Like when he compares banning AI in newsrooms to the early days of smartphones:

“Everyone said you need two phones, a BlackBerry for work, an iPhone for personal use. Surprise, surprise. Two or three years later, everyone’s got an iPhone and all their work stuff’s on it.”

His point? You can’t ban the future. People will bring AI in anyway.

Pete also watched the early AI disasters unfold, with newsrooms churning out what he calls “pink slime” or “AI slop” while trying to replace human writers with generative content at scale. He understands why journalists became one of the most skeptical professions when it comes to AI.

Pete is a journalist and AI strategist. Through his platform, The Media Co-Pilot, he helps newsrooms, PR teams, and creative professionals with practical advice and clear thinking.

His framework is elegant. Think of AI as belonging at the two ends of the journalistic process. At the beginning, research, idea generation, and refining your angle. At the end, repackaging your finished story for social media, newsletters, and different audiences. The middle, where the actual writing happens, stays yours.

“What AI does with a finished, human-written story, the sky’s the limit,” Pete explains.

To help prepare this episode’s introduction and transcript, we used Claude, an AI tool developed by Anthropic, under careful human supervision. Everything was fact-checked and polished by our team before publication. Transparency is important to us.

What we explore:

• The “two ends” strategy and where AI actually helps in the journalistic process
• Why banning AI tools is like trying to ban iPhones and what to do instead
• Using chatbots to learn what your audience really wants
• The niche revolution and why fewer eyeballs might mean a better business
• What most journalists get wrong when they try AI for the first time

Episode details:

Duration: 18 minutes
Guest: Pete Pachal, founder of The Media Co-Pilot
Host: Branislava Lovre
Format: Video podcast with full transcript

Branislava Lovre: Welcome to AImpactful. Artificial intelligence is reshaping how journalism works, from research and writing to publishing and audience engagement. But with all the hype, fear, and rapid change, many journalists are left asking, how do we work with AI in a way that protects our values. This episode is about that shift and how to face it with clarity, confidence and care. Our guest today has spent years at the intersection of journalism, technology and media strategy. Through his platform, “The Media Co-Pilot” he’s been helping journalists and editors navigate AI with a mix of practical advice, critical thinking and clear communication. Our guest is Pete Pachal.

Branislava Lovre: What inspired you to launch “The Media Co-Pilot”, and why did you feel it was important to focus on AI in journalism?

Pete Pachal: Well, I could tell AI was going to change this entire industry from when I first got my first taste of ChatGPT. And you know, there was obviously a big wave of articles about it. There were a lot of people publishing. I read an article just to show what it could do, and there was rightly a lot of consternation over what this was going to do with the media industry. So I was working at CoinDesk at the time. I was the chief of staff there, so I was broadly in charge of operations and strategy. So once ChatGPT was out there and I could see that media companies were adapting to it, I assembled a team internally at CoinDesk to essentially look at what this meant for our strategy and could we use it. Could we use it to sort of, do some kind of either content creation, or help our operations in some way. So I headed up the committee. We drafted a policy, and that’s still published on the site on how CoinDesk could use generative tools. That took a little while. And then we decided to start prototyping some use cases. So unfortunately, at CoinDesk, the crypto market wasn’t very kind to them. And even though we had a lot of ideas, there wasn’t really a lot of resources and will to invest in them. So after I’d left CoinDesk and sort of started thinking about what I wanted to do next, I knew I wanted to do something with AI, and I launched The Media Co-Pilot honestly to figure that out, because I didn’t quite know what it was, but I knew I wanted to study it and learn about it and pass on what I learned. So I started The Media Co-Pilot mostly as a newsletter, just to sort of teach myself AI and how it’s affecting media. Pretty shortly after that, someone reached out to me and said, Hey, can you come in to my team in our office here in New York, and walk our people through AI, and what they could do with it. And I was like, Yeah, sure, I could do that. And that went really well. So I thought, wow, you know, this is a valuable service. And within a short time after that, I started marketing that, like, hey, I’ll teach your creative team, whether it’s a newsroom or a PR team, or a marketing team, sort of the basics of using AI. And I got a lot of interest in that. I started teaching more regularly. I got on sort of teaching platforms, so people could buy the classes in various forms. And from there, I started doing some projects as well, for some of these companies. So what it’s evolved into is three things, since I do content, which is the newsletter and a podcast, I do teaching classes, courses. Some of them are generic, some of them are very tailored to the group and I do some consulting. So doing some projects for various creative teams, whether it’s a newsroom, a PR team or a marketing team. So that’s what I do now.

Branislava Lovre: What kinds of concerns are you hearing from journalists when it comes to AI in the newsroom?

Pete Pachal: I would say journalists are probably the most skeptical of what AI can do for them. And I think, that’s very understandable because of what has happened, and how the sort of media space initially reacted to generative AI because there was immediately, when ChatGPT came on the scene, a few, not that many, but a few major organizations thought they could use it to produce content at scale, essentially replace the idea of a human writing an article and, you know, creating what I think is probably widely regarded now as like pink-slime or AI slop, basically, you know, generative articles, on various subjects. And whether they’re human edited or not, this was sort of the most obvious use case, that’s also probably the use case for AI at this point, where it’s clearly not ready for the idea of content at scale. My view is that for almost all journalistic applications, it is a mirage. Now, there’s very specific, like really specific ways you could kind of do that and it was actually happening before ChatGPT. So you had AI essentially summarizing earnings reports, or sports, the results of a sports score, like a high school sports thing. And you could do sort of, all but really kind of a templated article, about that stuff. That’s very narrow. You know, that’s not – I wouldn’t even call that journalism to some extent. That is fine to do. But to actually take a more complex subject, try to sort of explain it in a more open-ended way and do that like essentially give you a finished product from your prompt. That’s not going to work on a mass scale. So understandably, journalists get really, really skeptical. You know, you’re just going to start cranking that slop instead of employing me. And it’s like, no, I think broadly we’ve moved beyond that. And when a newsroom comes to me, I think mostly they’ve gotten past that. They realize, No, we don’t want to do that. But you know what, there’s a lot of things we can do with AI and we’re probably underutilizing it.

Branislava Lovre: What advice would you give to journalists trying to integrate AI into their work without losing the integrity of their reporting?

Pete Pachal: I think that’s the main place you want to focus in terms of your use of AI are the two ends of the process. And I mean by that, like this story generation, like the initial research, the getting ideas, the refining of the ideas and getting as much AI in that process to get you good information for your story. And then you write the story. That’s probably the place you need to be the most careful about using any AI tool. Right. And I would say, like, for folks not comfortable, don’t do it. You know, like basically that’s that’s where like keep out, keep AI out of that process, to the extent that you can and still be productive. Right. So, like at a place like Semafor and they’ve thought it through and you have a process and it’s been vetted and tested fine. But if you don’t have that, I would I would stay out of that. But then on the other end of the pipe where you’ve done your story, again, AI can make a lot of, you know, a lot of utility out of what you’ve done by repackaging it, by giving you the headlines, the subheadlines, all those things that the article kind of needs to exist as a digital piece of content. You might even rewrite it in some formats for audiences. Now that gets, again, a little bit of diceyness. But because now you are an expert on the story itself and you’ve thought it through, it should be a much easier job of editing, say, a rewrite or a summarized version of that story so that, you know, like, oh, it’s hallucinated and put something in that’s weird, you’ll catch that in a much easier way. But, basically what the AI does with a finished human written story can be, you know, the sky’s the limit. It can repackage it for social and repackage it for a newsletter. It can rewrite it for a different audience, do all the things from the image to the headlines and everything else, and then sort of tie it with some automated systems to just sort of blast it all out. Right. Again, if there’s, anything generative, you’re going to want to vet at some level with human in the loop. But you know, you can do that in a much, much faster way now. And essentially get your stories out to way more eyeballs on way more networks than you ever could before.

Branislava Lovre: What ethical considerations come with using AI in journalism, especially when it comes to audience trust and transparency?

Pete Pachal: So, obviously be transparent about how you’re using AI, about how you are adapting to this new world in ways that are going to benefit the information you’re giving them. But also listen to your readers. And I think the new tools with AI give at least one major new avenue to be able to do that. And that’s chat, right. Like you can actually offer some kind of an AI feedback interface, whether it’s a general chatbot or something that is very focused on, say, your evergreen content or what have you. But basically giving that very easy way for someone to either respond to something or tell you what they want. So this is the thing like I don’t necessarily think chat is the great interface for everything, but I do think it is useful at some things. And right now I would say media companies are a bit challenged, like certainly smaller ones and sort of understanding exactly what their audience wants. Because frankly, a lot of media companies don’t really just have the money to have a big customer data platform and a huge marketing team to it, you know, which is kind of what you need to do to really understand your audience. But hey, if you have a chatbot and you put it strategically in the right places, so you don’t want, you don’t want to clip the experience, where it’s just bugging them all the time. Hey, talk to me, but if you do it in a smart way, they will tell you, what they want, right. They’re going to use that. And that is a fundamentally different search. You know, search gives you a little this gives you a more personal connection. And for the readers in your community that are very active and proactive and are essentially the most loyal, you’re going to learn so much from having them talk to you that you’re going to be able to then adjust your content strategy around that and then satisfy those loyal people even more and sort of deepen the loyalty. Right, now you still have the challenge of getting new readers, and that is going to be a challenge going forward with search, adapting to generative search and then social, kind of not still not being the greatest avenue. I think though, the idea that you are really serving a smaller group, really, really well is just something that is going to be fundamental to media strategy for the next several years. Essentially, going niche is going to be like basically your audience is going to be smaller. It will be, but that’s okay because you’re going to be deeper. You’re going to give so much value to those folks that you’re going to have the opportunity to offer subscriptions, to charge more for those subscriptions, to plan things like events, even if it’s just meetups and just do lots of tailored content and material for those people. And you’ll be able to make a business around that. And it could be something around the subject of expertise, it could be a region. But if you’re willing to do the work and use these tools to help you do the work you can, I think you could build something really, really good with fewer people and fewer eyeballs than before. So like, we’re going to basically redefine success as a media company, and I think it’s going to be all right.

Branislava Lovre: How do you imagine newsrooms working with AI in the near future? Do you see AI as more of a co-pilot, assistant, or something else entirely?

Pete Pachal: As they improve these models and it gets better at doing more stuff, I think slowly but surely AI became that companion in the newsroom you’re going to trust with more things. Do I think in the short term, it’s going to actually start writing stories in a way that humans do and sort of these inherently human things? No, there have to be some fundamental shift in its capabilities. Until we have that mythical AGI, where it’s just as good as a human even if it is, I don’t see the same spark, you know, just the way these systems work. And I’m not privy to what OpenAI has in its skunkworks or whatever. So, who knows, maybe they have a whole report of robots that could come out of it But I still think humans will be needed to sort of at least supply the spark of what they should be serving up. And then AI can sort of take on more and more help in sort of shaping that human idea. So that’s kind of what I think we’re seeing with the future of AI, I think with AI, because it’s not human and because failures can cascade and you can get amplified at scale, there needs to be a rigorous testing process, even at smaller places, to ensure that doesn’t happen. But I think, you know, if you don’t acknowledge that it will happen or that this companion can be very, very useful and will be a key component of your newsroom. I just think you’re really holding yourself back at this point. And, you know, you need to sort of start to get comfortable with it just sort of being there. But by virtue of it being there, it’s very important that all organizations, large and small, define what it can be used for, what it should be used for, because there are a lot of, as we’ve seen it, a lot of bad ways to use AI, right now. But you can’t ignore it

Branislava Lovre: Some newsrooms have gone as far as banning tools. What’s your take on that approach?

Pete Pachal: People will bring it in. It’s kind of like, what do you think about the beginning of the mobile revolution? Everyone was like, Oh, we got to have two separate phones and you got to have a BlackBerry and you got to super secure this one and your iPhones just for your personal use. Surprise, surprise. Two or three years later, everyone’s got an iPhone and all their work stuff’s on the iPhone. You just have to acknowledge the reality and work with it and, you know, put in some best practices around single sign-on and, you know, keeping your work Google account as separate as you can from your personal line, etc. And, you know, there’s ways to deal with the issue. But you can’t just stick your head in the sand and say, no, no, no, don’t use it. People are going to use it.

Branislava Lovre: And finally, what message would you give to journalists who are still unsure, hesitant, or even afraid of using AI in their work?

Pete Pachal: Just dive in, start using it. I know that’s kind of very default advice, you probably already have. Here’s the thing. If you’re a journalist, you’re hearing this, you’re like, oh, duh, thanks, dude. I wouldn’t have thought to try to use ChatGPT, but the thing is, you probably used it and I’ll bet you didn’t love the results. And I think you need to go back, you know, like if you haven’t used it in a while, one, the model’s a lot better and there are tools that you’re not using. I’m sure you’re underutilizing them. You probably don’t use memory. You probably don’t use custom instructions. Very simple things. You can look at, you know, some quick YouTube tutorials, to get you up to speed on that and then, you know, mess around with it, create a custom GPT to get some prompts that you know will help you write in your own style. I’ve got some ones I like. Again, get in touch if you want to know, I’ll be happy to give you one or two of those for free. But there are some things you can do beyond just write this thing for me, create this image for me that’s going to give you better stuff, right. And there are like I think with journalists in particular, it’s not just sort of create some writing, for me, it’s like the research and like curation of information is a big deal. Now AI do hallucinate. That’s an issue. But if you already have a certain amount of material and you have something you want to point it out, that’s a thing, that is relatively straightforward to do. It does take a little more time, but you will get good results much faster than, trying to go through all that stuff manually. So again, if this sounds like a complex thing, you’re not sure where to start. Feel free to email me. But again, there’s a lot of tools, there’s a lot of just even instructions on how to use out of the box, stuff like ChatGPT and Claude for that sort of thing. And you know, give it a go, go spend a couple of hours on it and try to always think about how do I genericize this for whatever story I have

Branislava Lovre: Pete, thank you for this conversation.

Pete Pachal: It was my pleasure. I loved going deep with you. Thanks so much.

Branislava Lovre: And to everyone watching or listening. Thank you for being with us. See you in the next episode.