Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

The Media Copilot

The Media Copilot artwork

Can you trust that clip? Storyful’s James Law on verification in the AI Era

Dec 19, 202543 minAudio
ShareXLinkedInFacebookWhatsAppEmail

AI has made verification a newsroom survival skill. In this episode of The Media Copilot, Pete Pachal talks with James Law, Editor-in-Chief of Storyful, the newsroom partner known for verifying the videos and social content that drive breaking news and internet culture. Law explains how verification has evolved from the early days of the ‘Arab Spring’ (a wave of protests across the Middle East and North Africa where user-generated social media footage first became central to breaking news), to today’s flood of viral clips, weather footage, and AI-generated video designed to look like real eyewitness content. The conversation gets practical fast. Law breaks down Storyful’s core verification workflow, why metadata still matters, and how the team pressure-tests every clip through three checks: date, location, and source. They also unpack why AI detection tools are not the silver bullet many newsrooms want, where these tools perform better, and why they fall apart when faces are not in the frame. They go further into the bigger trust problem: not just obvious deepfakes, but the “harmless” synthetic videos that quietly train audiences to doubt everything. The result is a world where authenticity becomes more valuable, not less, and where newsrooms may need to show their work to rebuild trust. A MESSAGE FROM PodPitch.com is a software that thousands of people use today to book podcasts with a 4% booking rate. It’s the most updated podcast email database and it comes with a custom-trained AI that learns YOUR voice and applies what works from more than 10 million previous pitches to optimize your own reply rate. Now one comms pro has the power of a 10-person team. Join Golin, Weber, Edelman, Finn, Broadhead, 5W and more in seeing a live demo today. Click here now to book time to check out PodPitch. Why this matters Verification used to be a specialized function. Now it touches almost every newsroom workflow. As AI-generated video becomes more sophisticated and easier to produce, the risk is not only that fake clips get published. The deeper issue is erosion of confidence in what people see online, even when the content is real. Law argues that newsrooms have an opportunity here: in a feed saturated with synthetic content, authentic reporting and authentic visuals carry more weight. He also makes a strong case for a cultural shift inside newsrooms. Audiences do not just want the story. They increasingly want the proof behind it: how it was verified, what sources were used, and what steps were taken to confirm it. What we cover * What Storyful does today, beyond breaking news * Storyful’s verification workflow: date, location, and source * How AI-generated video changed after Sora 2, and why it is harder to spot * Why metadata matters and why Storyful insists on receiving the raw file * The reality of AI detection tools and why they are not reliable enough for newsroom use * Why detection works better with faces and fails more often without them * The “delightful misinformation” problem and why it still harms trust * The “liar’s dividend” and how AI saturation makes people doubt real content * Why content provenance standards like C2PA face adoption challenges * A practical path forward for smaller newsrooms: training, verification habits, and embracing lo-fi video * Why authenticity consistently performs better than polished production on social platforms More about the guest LinkedIn: https://www.linkedin.com/in/jameslaw21/ X (Twitter): https://x.com/JournoLawJ Storyful leadership bio: https://storyful.com/about/ 📩 Enjoyed this episode? Subscribe to The Media Copilot on Substack, Apple Podcasts, Spotify, or your favorite app. On YouTube? Tap the Like button and Subscribe to the channel. For more AI tools and resources built for media professionals, visit MediaCopilot.ai. Produced by Pete Pachal and Executive Producer Michele MussoEdited by the Musso Media Team Music: “Favorite” by Alexander Nakarada, licensed under CC BY 4.0 © 2025 Musso Media. All rights reserved. © AnyWho Media 2025 This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit mediacopilot.substack.com/subscribe

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

I'm excited to tell you about a service I've personally used and had real success with.

SPEAKER_01It's called PodPitch. If you work in PR or communications, or if you're a founder or an independent creator trying to get your ideas in front of new audiences, this is worth paying attention to. PodPitch is a podcast pitching service that uses AI to help you find and book shows for you or your clients. There are tens of thousands of podcasts actively looking for guests. The hard part's always been finding the right ones and reaching them in a way that feels personal. Podpitch does that work for you. It indexes over three and a half million podcasts,

SPEAKER_01analyzes the fit based on your expertise, and then helps you pitch the shows that actually make sense for you. Here's how it works. You drop in your LinkedIn profile or your clients, choose your keywords, and PodPitch builds a smart media list and then writes tailored email pitches to each and every host. You can be as hands-on as you want. You can edit every message before it goes out or let the system do the heavy lifting. I've used Podpitch myself. I booked dozens of podcasts appearances all on my own without a PR rep simply by

SPEAKER_01using the platform. I can vouch that it works and it saves an enormous amount of time. If you want to get your voice out there in the fastest way possible, check out Podpitch. You can find a link in the show notes to book a live demo at podpitch.com slash media copilot.

SPEAKER_00You know, there was a time where newsrooms would say that a nightly bulletin would just show up and say, sources have told us this. And that was kind of good enough because the trust was there with the audience. But I think that that is no longer good enough and that audiences are craving the story behind the story.

SPEAKER_01Hello, and welcome to The Media Co-Pilot. It's a podcast that explores how AI is changing media, journalism, and content creation. I'm Pete Paschel, longtime tech journalist and founder of The Media Co-Pilot, and I'm excited to bring you fascinating conversations with fellow journalists, media executives, and the people building how the business of news will work in the future. This week, I'm joined by James Law, editor-in-chief of Storyful. If you don't know Storyful, it specializes in authenticating viral videos and social content,

SPEAKER_01basically confirming what's true before it spreads. If you've ever seen a news site post a viral video during a breaking story, there's a good chance James's team was behind verifying it. Obviously, the demand for understanding what's real and what's not is at an all-time high with AI in the mix. With tools like VO3 and Sora 2, there's just a lot of AI video out there. And it's not all Sam Altman cameos. James and I spend a good chunk of time talking about the increasing challenge of

SPEAKER_01being essentially a content authenticity cop in the age of AI. Not just what it takes to spot deepfakes, but also how do you even think about it now that everyone's attention is scattered and the media is just one of myriad places people are getting their information. James also shared his take on C2PA, which is an emerging standard to verify content automatically. And what other Providence standards mean? How are they working? What isn't working? And what progress needs to be made before the industry can

really start relying on these?

SPEAKER_01And we also zoomed out to the big picture. How does Storyful see its role in helping newsrooms ensure their content's legitimate, even as public trust in media is at an historic low? Well, I hope you enjoyed the conversation. And if you do, I'd be grateful if you leave a review for this podcast. Reviews, especially in Apple and Spotify, really do help the show. Please leave one if you can. Also, a quick note, you may not know this, but a big part of my business at the Media Co-Pilot is teaching AI to editorial teams. And I host a monthly virtual AI class for beginners.

SPEAKER_01It's only 60 minutes. Like everything I do, it's focused on practical takeaways for media professionals, people in journalism, PR, and communications. And it's only $49. Each one's a live class, so it's always up to date. And I guarantee you'll walk away with something you can apply to your day-to-day. You can sign up at mediacopilot.ai. All right. Now it's time for my conversation with James Law, editor-in-chief of Storyful. James, welcome to the Media Copilot.

SPEAKER_00Good to be with you, Pete. Thanks for having me.

SPEAKER_01Cool. Well, I'm glad you're here so that we can... you know, verify things a little better. I'm going to trust that this is you and not a deep fake version of you, but I guess I'm asking for that same trust on your end. So let's just assume we're both real right now, but I really want to dive into what you do at Storyful, the verification that you guys do with media and a lot of the facets around that, because it's obviously a pretty fascinating and vastly growing

thing in our media

SPEAKER_01ecosystem right now. Um, but I'd really love to start with, with you and your history at Storyful. So like what, what drew you to the company and, um, what is interesting about what we might call the verification space and then Storyful's mission.

SPEAKER_00Yeah. So, uh, I've been the editor in chief of Storyful for, for a couple of years now. And, um, Storyful is owned by News Corp, and I've got a long history with News Corp Australia. That's the accent you're picking up there. But I'm based in New York now, and this role appealed to me because ever since I was like a baby journalist, I very much enjoyed the myth-busting side of journalism. You know, there can be a lot of misconceptions out there in the world, and I always enjoyed kind of digging into that and kind of

educating the audiences

SPEAKER_00about things through that lens of fact and truth. You're the well-actually guy. Well, actually, yes, indeed. There's a satisfaction to doing that. And that's the core. No, no. But that's the core of what Storyful does. It's all about verifying what's true, what we know for sure, and getting that out to other newsrooms. And it's all also about like building trust with audiences,

SPEAKER_00which is very much a challenge for newsrooms at the moment. So that's something that we can really help with.

SPEAKER_01Cool. Can you give me a sense of the scale of the company right now? And also like what a typical day or even a workflow is like, because, you know, I've been tracking Storyful for a while. I remember we were using it quite a bit back when I was at Mashable, which is like 10 years ago now. So give me a sense of like, you know, obviously lots of stuff has happened since then. That was when social was huge. I mean, not that social isn't huge anymore, but it just feels like the scale of things that are being produced and being circulated and being touted is like either newsworthy or helpful is just vastly increased.

SPEAKER_01So, you know, give me a sense of like how much you guys handle it in a typical day and then how you handle it. Well, to your point, when Storyful began, it was very much focused on breaking news. It was inspired by the Arab Spring and a lot of the kind of big news that was

SPEAKER_00coming out of that news event was happening on social. And there was a real need to verify that video before it got published. And so that was the kind of beginning of Storyful. But as you identify over the 15 years we've been in existence, we very much have expanded what we do. Newsrooms don't just want breaking news. They also want weather coverage. They also want the lighter side of what's happening on the internet. And we're very much experts at that as well. You know, all of the viral content that very much feeds engagement from

audiences.

SPEAKER_00We're also quite adept at that content as well. And then beyond that, we also at Storyful have turned our expertise as, you know, We sort of know social media better than anyone. We've also turned that into an intelligence business that helps give businesses insights around what people are saying online, conversations that are happening and shifting online, and how brands can kind of navigate that in order to protect their reputation. So certainly there's a lot of strings to our bow these days.

SPEAKER_01So is it kind of a little bit of just finding where the origin of certain conversations comes from? And when that happens and you're kind of finding about these conversations, whether it's for publishers or for brands, is there some role you play in the debunking or redirecting, I guess, of those conversations? Or do you sort of leave that up to them?

SPEAKER_00Uh, so in the intelligence side of the business, we are just kind of really trying to get a good sense of what people are saying and how conversations might be shifting, where there are risks for brands and things like that. So less about, uh, demunking in that space, but very much in the editorial space, which I look after, we do see that as an important role we have to play in, you know, pointing out when there are, there is misinformation proliferating online or trends that we're seeing, um, and I'm sure we'll get in to talk about AI in more detail,

but we certainly do try to actively share with our newsroom partners when we see a

SPEAKER_00piece of misinformation that's going out there. For example, with the advent of the Hurricane Melissa in recent weeks, the big hurricane that hit Jamaica and also parts of the US, we started to see a lot of AI videos circulating around that event. So we kind of collated all the fake ones together and shared them with our newsroom partners to let them know that they were fakes. And we do that a lot. We also see a lot of times when breaking news is happening,

old videos will resurface and be republished with the intention to mislead people,

SPEAKER_00such as in some of the LA riots earlier in the year that all the videos were being republished to try to trick newsrooms. So when we figure that out, we share that with newsrooms to debunk it.

SPEAKER_01And, you know, we're obviously starting to talk about AI a bit, and I think it's a good transition on, like, first speak to sort of the scale of it that you've seen over, I guess, the past couple of years, just in that period of how much AI has changed in terms of like what the, not just the sort of volume of stuff, but also like the quality, like how has it made your job, I guess, more, I'm not sure if you want to say difficult or interesting or both or something else. but describe what the effect has been.

SPEAKER_00Storyful are kind of the experts at picking this stuff out. So, you know, in the sort of two years I've been in this role up until today, we haven't been fooled, touch wood. And our journalists are expert enough that they were able to pick it with the naked eye. And there was something about AI in, you know, say a year ago where it had a sheen to it. There was something about it that looked a little uncanny. And because our team is so used to looking at the real stuff,

they were quite good at picking it out before it even really got too far along our

SPEAKER_00processes. I will say, however, that that has changed. So since the advent of Sora 2, it was kind of this new AI generative model was launched in September. We've seen a big shift in the sophistication of these deep fakes. the ability to create something that looks more like user generated content, which is what we trade in, has very much advanced. There's been a quantum leap forward. And,

SPEAKER_00you know, I still think that Storyful is very well placed to protect against this content and, you know, very well placed not to be fooled. And we see ourselves as leaders in this space and are really trying to help our newsroom partners, giving them tips about how to look out for it and how to combat it. So we are seeing it advancing, um, quite rapidly, but we still feel very well placed to, to help the industry, uh, navigating it.

SPEAKER_01So you said sort of knock on wood. So it sounds like you've, like you said, you haven't been fooled yet. Is there been, it sounds like it's getting closer and closer. Like maybe there's more close calls, uh, lately just because people are getting more sophisticated with tools that are getting more sophisticated. Um, you know, what is your, um, What is your outlook for the future here in terms of when, you know, can can you guys keep up, I guess, is the concern, not just as necessarily with the scale of it, which we didn't really speak to,

SPEAKER_01but I presume there's just lots more of this stuff just circulating now. And but with the like, how close are we to that? I'm not sure what the term is. It's not quite uncanny value. Maybe it's on the other side of the uncanny valley, right? Because we can all sort of still see that sheen that you talk about, but it is getting better and better. And I do find I'm getting almost fooled in the sense that I'm kind of drawn in and I'm no expert. I obviously follow this stuff, but I'll need to watch it for usually like 10 seconds in and then something unreal happens and it's like,

SPEAKER_01oh, okay. But other than that, it's tough.

SPEAKER_00Yeah, well, to answer your question, we remain very confident that the checks and standards we have at Storyful are going to be strong enough to ensure that we are continuing to verify things to a very high standard. And to give you a bit of a sense of what they are, With every video we have, we check, we have three essential checks. We check the date, we check the location and we check the source. And usually if it is an AI generated video,

SPEAKER_00something coming from thin air along that process, it should get tripped up. So let me run you through it in detail. So there's the date. So if something happened on November five, There should be other corroborating information. Let's say there was an explosion here in Sixth Avenue, right? And our London team was looking after that story. What would they do? If there was that big explosion, surely the fire department would have something to say about it and be able to

SPEAKER_00confirm that date. We'd also be wanting to check geolocation. So we do that with every video. And essentially what that means is we're matching what's happening in the background of a video, the buildings, the landscape, the signs, to open source maps. So Google Maps, for example, that would allow us to kind of match what we're seeing to a real location. AI sort of still can't really do that convincingly. And then the third thing and probably most crucial is the source.

SPEAKER_00So say we saw a video we were interested in, we would want to find who the original uploader of that video was. We'd oftentimes want to get them on the phone. We'd want to check their address. We'd want to make sure that if it was purported to have happened in Dallas, is this person from Dallas? Do they have a habit of posting newsworthy videos from Dallas? And so along all of those checks, we would feel quite confident that they are still strong enough to ensure that

SPEAKER_00we're not publishing anything that is completely fabricated. And to what extent do you use AI in terms of whether it's the filtering process or

SPEAKER_01just highlighting the things that are in need of verification, like What is there? Is there an AI layer to what you're doing?

SPEAKER_00Yeah, we're certainly... We have a dual approach to AI. One of them is, and the core of what we're trying to do is help our newsrooms not publish anything that is fake. And that's kind of job one. But we also see AI as having a big opportunity for us. And we're in a very big experimentation phase in our newsroom about figuring out ways that can help us be more efficient, more effective and move faster. So some of the things we're exploring at the moment is whether

it can help us

SPEAKER_00surface newsworthy videos faster, whether it can help us in some of our kind of back-end processes that remain a bit manual, could that help us speed up some processes? Because if we license a video, there's a contract and how do we get that to work as efficiently as possible? We see it as a big opportunity and, you know, you need to remain nimble. You need to embrace the new technology and that's where we can see it having a lot of benefit for us.

SPEAKER_01And is it done in some ways to detect some of the things, whether they're artifacts or not, that humans might not be able to? And, you know, that's kind of what I was alluding to earlier. Like I was talking a lot about human intuition and eyeballs, but maybe that's over time just simply going to get less reliable, right? Just as more sophisticated these things do. And you're going to need to rely even more on machine authentication or assistance with the authentication, I should say. And is that sort of an equation or a balance that you guys are

seeing that is sort

SPEAKER_01of shifting in the direction I'm predicting?

SPEAKER_00So it's hard to say what'll happen in three years with this technology. But for today, Storyful's done an analysis of some of the detection tools that are out there. We looked at about seven of the ones that are out there available to test and on the market. And what we have found is they have a role to play, but they are very far from being effective. we would probably rate it about 70% that it can actually detect

manipulations.

SPEAKER_00Well, let me give you our analysis in more detail, the caveats. And I'll start that really newsrooms want 100%. And so without that, you need to really question whether they are... I think people want a silver bullet. And our analysis is that they are not a silver bullet, that they perhaps have a good role to play, but that you continue to need that human expertise to ensure that you don't get caught out. But what we found with the tools is they can work to about 70%

accuracy when there

SPEAKER_00are faces in the video. So they are good at detecting human facial movements, matching the sound of speech to the movement of lips. It's getting quite good at that. But when a human face does not appear, its accuracy dips to below 50%. So if there are animals in the video, it really can't quite understand how a bunny works and jumps around accurately. It has issues with landscapes.

It's not really able to detect all of the different types of video that are out there.

SPEAKER_00Another thing that it's... Go on.

SPEAKER_01Oh, I was just thinking like what might also trip it up is just simple editing, which is to say that if you think about it, you know, Often when you want to check provenance, you're detecting whether or not this is fabricated at all, or if it truly represents the event. But sometimes you have something like it's a concert and someone put on a filter and, you know, there's just kind of stuff that looks, makes it look a little unreal, but it's actually a real event. And that must be a bit of a challenge for these systems to sort

of discern.

SPEAKER_01Cause it's like my flag, it was like, oh, this is as telltale signs of being fake, but it's like, wow, it's, it's really just been edited a little bit.

SPEAKER_00Precisely, or yeah, there's a video screen in the video and then it thinks everything in it is fake. So that's right. It does struggle with that. And I mean, the other thing is newsrooms when, so say a newsroom is interested in using one of these, If you found a video on X, for example, and you want to test if it's real, if you put it into one of these tools, you don't have the video with the metadata because what you're putting into the

SPEAKER_00system is a duplicate, is a reproduction of that, really. And so without the metadata, the original file, it can also, again, struggle to really make a good determination. story full on the other hand if we license a video we insist that the person who recorded it sends us the raw file and that's hugely important in the verification process it's going to be a big file right that starts to give you some confidence

we can look at the metadata of that video and test whether it was filmed on the

SPEAKER_00date they said in the location they said tick we can test whether that um what what is in the metadata of that file matches with where this person says they live so you know all of this is to say that i think the tools have a part to play and it'll be interesting to see how they get better but you still will need that human expertise to to give you that confidence going forward the other challenge for the tools is that as the uh

SPEAKER_00the generative models improve, the tools have to improve in tandem, and they're often a little bit behind. And so ensuring that they are across all of the huge advancements happening in Sora 2, for example, is a big challenge for those tools. So we're watching this space. We hope that they will improve, but that's our analysis as of today.

SPEAKER_01Well, I'm glad you brought up metadata because it brings me, gives me a good opening to get to the technical parts of the discussion I wanted to bring up with you, particularly with markers, content provenance and storage standards, I guess they're emerging standards like C2PA. and i wanted to get your thoughts on um just generally how um how useful those new standards are uh today as they're being implemented but also um

what is holding them back from really taking on a greater role in this whole process whether it's

SPEAKER_01story full or just even more broadly Um, cause I definitely get the sense and you, you tell me your take, but that it's, it's slow going that it's, there's still a lot of sort of education platform adoption, et cetera, to really get these sort of, I guess what you would call maybe more automatic. and, you know, technical handshake verification to actually work or at least be effective. Your thoughts on C2BA, the whole sort of background of technical content provenance

and what might need to

SPEAKER_01happen for that to take hold and how useful it would be ultimately.

SPEAKER_00Look, I think it's a very noble aim. And I think that, you know, we're pleased to see that there is people trying to solve this problem. But our view at Storyful is it's hugely challenging. So I think there's different places where it applies differently. So for professional photographers or professional videographers, allowing them to kind of watermark their content through the process has a lot of value and should be pursued.

The challenge for us at Storyful is we don't really deal with professionals. The content that we acquire and that we're seeking every day is shot by everyday people who probably just once or twice in their life happened to be there when

SPEAKER_00something newsworthy happens and they filmed it with their phone. Without Apple's endorsement of this and it trying to place that watermark at the very beginning of this process on every iPhone, it's very difficult to see how it can play out through the whole life cycle of a user-generated video. And I see no impression that Apple is embracing this concept at this stage. So that's the challenge for us in trying to fully embrace it is

that without

SPEAKER_00everyday people having that in everything they film, it won't work in the way that we do. So it's a big challenge. And for professionals as well, it kind of needs everyone's buy-in. to really work. So it's quite a challenge. I think it's a noble aim, but it is a challenging prospect.

SPEAKER_01Yeah. Like you say, it kind of has to be end to end, like you need obviously the device makers, like you said, and certainly the ones that are, um, most fueling this type of content. Um, you need to start there, but also like, I think there's a big platform question, right. If like you mentioned X, there's obviously tons of other networks like Facebook, TikTok, et cetera. And. without some kind of participation or marker or whatever, I just feel like you're always going to be relying on people

all the way down.

SPEAKER_01And even with all this participation, you're going to want people. But to the extent that you don't really know anything, going into just first seeing something because this doesn't exist just makes it way more challenging. Maybe that's good for you guys because maybe there's more demand for services like Storyful. But ultimately, someone's got to blink first. And to your point, I don't really see anyone really on the big tech side of it blinking.

SPEAKER_00No, and I don't know that they have an appetite for it at this stage. I think that their priorities have shifted quite significantly where this is not a huge priority for them. So I don't see any of them being right on the forefront of this. And it also must be said that Looking at SOSORA 2, for example, it obviously has the logo or the watermark on its content as it goes out, but they are so easy to remove.

SPEAKER_00One would think that the C2PA, that watermark as well would need to be very sophisticated to ensure it can't just be easily removed by anyone with a tool. So that's another one of the challenges.

SPEAKER_01You bring up a good point because, you know, the way I characterize it there, maybe nobody's doing anything. But to your point, there's the watermarks, which are a fairly common thing, particularly for free services. YouTube has things for creators when you upload. You have to sort of market. They're basically the terms of services. You have to mark something that has been AI, like essentially a deep fake. You have to mark it as such. And even other, there's other examples I could cite, but there's things that are sort of this ad hoc,

SPEAKER_01almost lip service to it. It's probably more than lip service, but I guess it's relying a lot on like self-policing, I guess, and people being good actors and having the weight of either a single platform and potentially getting thrown off. as some kind of incentive here. Do you feel that's doing anything or is it just woefully inadequate?

SPEAKER_00It is inadequate. It isn't adequate, to be clear. And I think that that is very much the way that the platforms have approached a lot of these issues around, yeah, provenance and around fact-checking. Over time, they... I think it's clear from their behavior that they feel they are not the arbiters of truth. They are simply a platform and that they just allow people to upload content or say what they like. And that that's,

SPEAKER_00you know, I think they put it onto the community to police itself has often been their approach. I think that is inadequate. I think that they have much more of a responsibility to, uh, to influence what is on their platforms and to kind of uh they have a responsibility to ensure that there isn't misinformation out there and that there isn't harmful content on there and uh i think there is more to do and more to do in

protecting copyright and and ensuring that misinformation doesn't proliferate

SPEAKER_01Yeah, that's fair. And I also, to sort of get maybe zoom out a bit on the bigger issue, because, you know, to be fair to the platforms, you know, they obviously need to provide a forum for people who are doing, say, satire, right? And, you know, like, in other words, it's not like there's no legitimate use cases for AI videos. Obviously, there's a lot. And You know, it's a question of how much good faith are you doing it with

for your audience? Are you, you know, are you literally trying to fool people with a deep fake, et cetera?

SPEAKER_01And that obviously so. So in other words, like there needs to be. And I think that's sort of to the point of the YouTube stuff I was talking about earlier. It's trying to sort of serve both audiences. um i guess i would sort of question almost in a bit maybe a devil's advocate kind of way that as these deep fakes as we've seen um go out there when say someone makes some politician say something they never did and that got circulated um it's also like because it's circulated pretty quickly the way our social networks work

it's also it seems to be debunked fairly quickly like in other words just the way

SPEAKER_01our society works if something of a fake video gets spread you're not going to be talking about it a week later or a few days later like it's real. And maybe that's me because I'm in information and AI bubble and I keep track of stuff. So maybe that's naive of me to say. But I do feel like there were fears about deep fakes that were circulated probably a couple of years ago when all this technology is relatively new, which haven't sort of...

SPEAKER_01turned our, like, it hasn't turned our elections, we just had one here, like in November, that into some kind of circus where everyone's, I don't even know what's real, right? So, you know, again, I think this goes beyond, like, just sort of what Storyful does and, you know, the greater sort of ecosystem of, like, what a deep fake is and then what the effect it has. But what is your take on that in terms of like, is the system somehow self policing itself or what are we missing here?

SPEAKER_01Is it the wrong question?

SPEAKER_00So I think in terms of like outright bad actor content, so like making a candidate say something they didn't, I agree that the community has been pretty good at debunking those and we have a role to play in that and there's plenty of others who do. So you're right that there hasn't been that kind of outright misinformation that has swayed an election or caused a massive headache. what we find is it's a little bit more subtle than that it's

almost like the the proliferation of the delightful uh misinformation is is almost a bigger problem right so the videos right where they create what story for had a video where we verified uh some coyotes jumping on a trampoline and it was like a night vision

SPEAKER_00camera of that and 100% real, verified by us, distributed around. And that almost created a trend of others creating bunnies on a trampoline, kangaroos on a trampoline. And they actually fooled people. People thought they were real. They shared them around and only found out maybe later or maybe never that they weren't real. And the way that that is insidious is like any one of those videos on their own is trivial, is probably funny,

SPEAKER_00and people are like, oh, whoops, I shared AI. But what it does at scale, if this content saturates our feeds, there's a real risk there. We already know that... audience trust in what they see online and certainly in what they hear from news brands is at an all-time low and so i really do think we are at this place where if if this content proliferates even further it does add to that impression that i can't believe anything i hear or see and certainly with younger

people who have no automatic trust in any news institutions really that becomes a real risk and certainly if newsrooms republish this stuff it can even worsen that impression

SPEAKER_00of this lack of trust so that's where i think it is problematic

SPEAKER_01Yeah, I agree. And I think there's a, you know, term that gets thrown around sometimes called like the liar's dividend, which I think is sort of the, at least part of what you're talking about, which is to say, once this stuff is all in the system and it is, you have to look askance at everything and even say something that does come from a new source that is verified through your service. People will just be going like, oh, it's probably AI because I choose not to believe it because I just see AI stuff everywhere.

SPEAKER_00Right. Yeah.

SPEAKER_01Again, I don't know if that's something you guys can solve, but I mean, it's something you kind of like almost need to find your place in this system. You know what I mean? Because, you know, it used to be 30 years ago, of course, the media was the media and they were the gatekeepers. And if you saw something generally on one of these highly circulated publications, you could trust it. And now to your point, media is much less trusted. It's much less part of our broader attention. You know, you might be spending a ton of time just on TikTok seeing AI video after AI video. And then the, the actual news, the verified stuff is a much smaller portion of what you're seeing.

So it's just this, it's almost like we're kind of halfway floating in unreality anyway.

SPEAKER_01Um, and you just, the, the, the real stuff now is cheapened, you know, and I, I, again, I'm not quite sure how to solve for that and I'm sure you're doing your best, but like, what is story falls place and sort of trying to, I guess, increase that. well i i think that there is an opportunity for newsrooms here because although the content is proliferating it's not clear how valuable it is what is the value of that you know diverting bunny video if it's not real i think

when people see them or if they get fooled by them they feel cheated and they feel let down if they're

SPEAKER_00not properly labeled right So in this environment, I think authentic content has more value, in fact. And I think that the way that newsrooms can respond to this problem is by... breaking down some of the barriers and reveal, almost like revealing their homework to the audience is a way we can build up that trust and increase the value of the reel. So, you know,

there was a time where newsrooms would say that a nightly bulletin would just show up and say sources have told us this and that that was kind of good enough

SPEAKER_00because the trust was there with the audience. But I think that that is no longer good enough and that audiences are craving the story behind the story. Right. So there's the there's the reporting. But then there's also like the side story about this is how we know it. This is how we went about finding this out. These are this is who we spoke to. This is what you know, sources we drew from. I think audiences are very interested in that.

And I think that's the opportunity for us in shining a light on the hard work we do and trying to bring the audience into the process so they can understand its value.

SPEAKER_00Because the value of the kind of viral videos that we do is that they are unexpected and delightful and real and unscripted. And I think that that will always have more value than something that has been designed to delight you. It's the randomness of it that makes it valuable. So for smaller newsrooms and outlets that obviously want to be part of the solution

SPEAKER_01and not the problem, but don't necessarily have resources to verify everything, or if they don't have it, maybe they just don't publish anything. I feel like there's a... There's a couple of different forces here that are, they're kind of between almost a rock and a hard place. One is that people are demanding more visual content. It's just sort of part of our diet. Everybody wants to be spending more time on these networks that are, we mentioned a bunch of them already.

And I think most publications and people wanting to reach an audience sort of need

SPEAKER_01to play there, but doing good visual content takes resource. or getting stuff from you guys even. And so what is the sort of route forward for them to either both verify, but also still wanna play in these spaces with trustworthy content? What are some key first steps to publications that wanna say like, okay, we wanna do more stuff to entice our readership visually, but we only have so many resources and so many people to go around.

SPEAKER_00Yeah, and look, Storyful is very connected to the industry. We have many clients, so we sort of know how straightened the times are and how journalists are being asked to do much more with much less and, okay, you've written the story, now do the video, now do this, now do this. So we're very sympathetic to those pressures that every newsroom is under. But I would say that the kind of two things you can do, I think it's about continual training.

SPEAKER_00So, you know, Storyful has a role to play in this. We're out there sharing tips on LinkedIn and trying to, you know, help the newsroom to skill up in this space. So, ensuring that the staff is continually trained around what are the clues to look out for, for AI video and all that kind of stuff is one step. And the other thing I would say about video, I know that smaller newsrooms can struggle to get a video strategy or to how do we

SPEAKER_00do, we've done the text report, how do we turn that into a video? I think that newsrooms need to embrace the lo-fi nature of what social video actually is. A lot of the content that we acquire, a lot of the people who are successful on TikTok, for example, they don't have studios. They don't have training in how to read an autocue. They are simply people putting their phone in front of their

face and talking like

SPEAKER_00I'm talking to you now. And younger journalists are perhaps less afraid of that than people think. And I think encouraging experimentation, encouraging, you know, talk to your phone for a minute, explain your story like I'm talking to you in the bar and put it up on your socials and just see how it goes because it doesn't need to be highly produced. It's better if it's not. So just have a play around and I think the audience will come

along with you.

SPEAKER_01Nice. A plea for authenticity. Authenticity is all you need almost if you were to author that paper.

SPEAKER_00Indeed. Indeed. It's hugely important. Like on TikTok, people do not want the polished performance. They want the real person. And encourage journalists to be themselves and to show their personality. If they're a huge nerd about a certain subject, be that nerd online. People will resonate with that. That'll find an audience.

SPEAKER_01Definitely worth a lot more than a deepfake. James, this has been great. We'll leave it there. I really appreciate the time.

SPEAKER_00Great to chat to you, Pete. Thanks a lot.

SPEAKER_01Thanks for listening to my conversation with James Law of Storyful. If you enjoyed this episode, I'd appreciate if you'd take just a minute to leave a rating and review for the Media Co-Pilot in your podcast app. Also, don't forget to subscribe to the Media Copilot newsletter and check out mediacopilot.ai for our classes on AI for journalists, PR professionals, and communicators. Seeing you in the future.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Why Yahoo still matters and what it knows about the Future of News

    Jan 16, 202641 minAudio

    For years, Yahoo News has occupied a peculiar position in the media ecosystem. It is not a social network, not a traditional publisher, and not a search engine in the modern sense. Yet millions of people still use it as...

    Credit the people on this episode

  2. Best of the Year: Inside the AI shift that's transforming media and journalism

    Dec 28, 202531 minAudio

    This special episode of The Media Copilot is a look back at the conversations that defined the past year… the questions, tensions, and turning points shaping how media, journalism, and technology intersect right now....

    Credit the people on this episode

  3. Practical AI in the newsroom with Darla Cameron of The Texas Tribune

    Dec 12, 202541 minAudio

    AI in journalism can feel abstract until you talk to the people actually shipping products inside newsrooms. In this episode of The Media Copilot podcast, host Pete Pachal talks with Darla Cameron, Chief Product Officer...

    Credit the people on this episode

  4. Building the next era of the open web with Adam Singolda

    Dec 5, 202543 minAudio

    AI is rewriting the rules of digital media and few people have had a closer view of the shift than Adam Singolda. In this episode of The Media Copilot, host Pete Pachal has a candid conversation with the Taboola founder...

    Credit the people on this episode

  5. Inside Time’s AI push: Mark Howard on building an agent, not just another widget

    Nov 21, 202551 minAudio

    If you grew up with Time magazines lying around the house, you probably have a sense of memory of that red border: serious, premium, grown-up news features - not just the facts. Now imagine that same publication trying...

    Credit the people on this episode