How Ad Dollars, And Some AI, Might Restore Our Shared Truth — With Vanessa Otero

Vanessa Otero is the CEO of Ad Fontes Media. The company rates publications based on their biases, and allows advertisers to concentrate their spending in news media that may disagree, but isn't so wildly biased it loses rooting in reality. Listen for a conversation about the ad industry's broad defunding of news, what it would take to return that money, and why artificial intelligence might help scale the efforts of Ad Fontes' human news raters. --- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. For weekly updates on the show, sign up for the pod newsletter on LinkedIn: https://www.linkedin.com/newsletters/6901970121829801984/ Questions? Feedback? Write to: [email protected] Learn more about your ad choices. Visit megaphone.fm/adchoices

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Alex Kantrowitz?

  An advertising CEO who wants to restore our shared truth by directing ad dollars to publications with a reasonable amount of bias joins us right after this. This episode is brought to you by Genesis. What does it actually take to put a Gentik AI to work across an entire enterprise? I recently spoke with Genesis chairman and CEO Tony Bates about why orchestration is emerging as a competitive advantage in customer experience and what companies need to get right in the AI era. Then I sat down with Adam Mitchell, head of enterprise business solutions at Voya Financial to go inside a large scale agentic CX transformation, including the business decisions, trade offs, and lessons along the way.

• 0:43Alex Kantrowitz?

  You can watch both conversations now on my YouTube channel today. It's never been more challenging to get frontier compute. How do you know where to look, who to trust, or how much to pay? Akash is an open cloud that provides easy access, flexibility, and peace of mind. That's how companies like Razor were able to cut their inference cost by 15 times. Visit akash.network/alex for a free consultation. That's akash.network/alex for a free consultation.

• 1:15Speaker 01

  LinkedIn presents.

• 1:20Alex Kantrowitz?

  Welcome to Big Technology Podcast, a show for cool headed nuanced conversation of the tech world and beyond. We're joined by a great guest today. Vanessa Otero is the CEO of Ad Fontes Media, which rates publications on how they're biased and lets advertisers buy ads based on those ratings. It's one of the more fascinating companies I've come across in my years as a reporter, and I'm so excited to bring you this conversation today. Vanessa, welcome to the show.

• 1:45Speaker 01

  Hey, thanks for having me, Alex.

• 1:47Alex Kantrowitz?

  Thanks for being here. Let's start here. It's clear that our shared truth has kind of dissolved, right? There are people in the US and all over the world who can look at the same news story and either look at it completely different with a different set of facts, or maybe even some folks on one side of the political aisle won't even know that a story is happening while another side is going to be fully tuned into it. So I'm curious, from your perspective, how did you initially realize that this was a problem because you're trying to solve it? And what can we do about it?

• 2:17Speaker 01

  Well, this all started for me back in the 2016 election. And the way that people would fight about news and politics on Facebook in particular, back when the, there was lots of news shared on the news feed, people would share really biased and unreliable sources with each other. And I noticed that people couldn't convince each other of their points, right? Like, people would just completely discard the media from the other side. And it appeared to be dysfunction of the fact that we had so many more news and information sources than we ever had before. So it was really easy for people to get into their own little filter bubbles.

• 3:02Speaker 01

  I mean, in the years right before that, that's when that term was really coined. And like you said, there's no shared truth. I noticed that folks would, they'd have divisions because of both polarization, but also because of, you know, people broadly call misinformation. Like, it makes sense that people on the left and right, political divides, you don't see eye to eye because they don't see, you know, read and hear the same things. But misinformation is like that, too. Like, if you have one set of folks that believes a certain set of facts and others just don't believe those facts, that creates division in that way as well.

• 3:43Speaker 01

  So, you know, what this all started out as was, you know, an infographic just to plot out a few dozen of the news sources that are out there to show that there's a difference between stuff that's, you know, highly reliable, stuff that's okay, stuff that's really problematic, things that are left and right, and things that are even more so.

• 4:04Alex Kantrowitz?

  So I am kind of curious, like, okay, these are big things that you're going to tackle, bias and misinformation. Like, how do you even begin to consider, like, doing this in a way that's not politically motivated?

• 4:16Speaker 01

  The answer is you have to get super granular. So for us, that means human analysts from left, right, and center looking at individual articles, individual episodes, individual, uh, your TV and podcast content in order to, like, look at sentences, headlines, graphics, all the things that make it up to come to a rating on that.

• 4:41Alex Kantrowitz?

  You know, maybe the idea is, and I think this is what your idea is, right?

• 4:45Alex Kantrowitz?

  You know, don't eliminate the idea of biased publications or publications that reflect one political standing or another, but sort of concentrate the ad spend into those that still are, you know, that believe in the shared truth versus are extremely skewed. So is that what you're trying to do? And how does that happen?

• 5:05Speaker 01

  So when we first approached ad tech companies and brands and publishers, like the whole idea was, you know, we want to help it change the incentives in the media ecosystem. And we thought that the the brand's biggest use for our data would be to help them avoid extremely polarizing and misleading and inaccurate content because it's bad for society and also it's probably bad for the brand, right? But when we started approaching more and more brands, we we found so many of them would tell us like, well, that's not really a problem for us because we just don't advertise in news anymore at all. And

• 5:47Speaker 01

  yeah, that to me, that seemed that seemed like worse, right? Because the you have this worst of all possible worlds where the misleading, highly polarizing content, it's fairly cheap to produce, right? You don't need, you can have a person or five with a website and you can churn out just rage clickbait material pretty easily and get enough like programmatic ad dollars to stay afloat and you're great. But then the high quality investigative reporting, the editorial newsroom stuff, that's expensive. So if you have enough brands pulling away from that kind of content, then you see this degradation in our media landscape where

• 6:30Speaker 01

  good journalism suffers and the sort of garbage out there thrives. So that's what we're really trying to change in the ecosystem.

• 6:40Alex Kantrowitz?

  Right, and you read a, you wrote a op ed recently in Pointer, and you said this one sentence that to me really stuck out, which is that brand safety, this rush toward brand safety is misguided. I mean, it seems to me logical that brands would want to put their ads on content that is not controversial. So why do you think it's misguided?

• 7:00Speaker 01

  Yeah, so the brand safety as an industry only really got started, you know, within the last, you know, 10 or 12 years. And it was, I think the first big moments in the, in this brand safety movement were around YouTube videos where there are terrorists beheading people and ads were showing up next to that kind of content, like this user generated content. So your companies rushed in to provide solutions around, oh, let's not show up next to things that are violent. Let's not show up next to things that are sexually explicit. Let's try to screen out, you know, bad words that, um, and do

• 7:39Speaker 01

  this at scale. And it sort of grew and grew and grew to this place where it's like, anything that's, you know, kind of yucky at all, um, brands are like, okay, though that, let's, let's stay away from that. Let's not put our advertising dollars next to it. And that came to encompass news. However, there's not great data, there's not really any data that shows that, uh, brands suffer any sort of, you know, backlash from consumers if they're advertising in high quality news. Examples I give, you know, Wall Street Journal, uh, you know, a few, a few months ago, they had

• 8:20Speaker 01

  a, an expose about Instagram and how, you know, um, sexually explicit content with minors was showing up on Instagram. And it was like this investigative reporting that was, you know, pretty important. And in the Wall Street Journal, you know, they have embedded ads from lots of big brands, right? And there was no backlash to those brands for appearing adjacent next to content that was reporting about, um, this unpleasant subject, right? No one said, oh, that, that advertiser, they support, you know, child porn on Instagram. No one, no one says that, right? But the brand safety tools that exist to

• 9:02Speaker 01

  You keep folks away from from news. They're so broad, like so many brands use keyword blocking lists and category exclusions in in order to stay away from anything that's quote controversial or quote negative. And it's just swung too much in that direction that it's really hurting news publications.

• 9:25Alex Kantrowitz?

  And you say that more than 80% of advertising has evaporated from the news business since 2005 as print has collapsed. So how much of this money moving away from the news business is due to the collapse of print? And how much of it is due to the fact that advertisers just don't want to be on news stories anymore?

• 9:43Speaker 01

  Well, it's a combination of a lot of things, right? A lot of things have changed in the news and technology landscape since 2005. I mean, you have the rise of, you know, Google and Facebook. I mean, so much spend that was going towards news moved over to search and these walled gardens, right? So it's not just that brands have moved away from news, but the brand safety movement has has certainly not helped. So and it's ironic because in, you know, social media platforms, brand safety is very opaque. There are really no guarantees on it. So it's ironic that brands would say, well, we don't want to advertise next to news because

• 10:28Speaker 01

  it's not safe, yet they'll just put that money in Facebook, X, you know, YouTube, where there's really not a lot of insight for those brands into the, you know, brand safety on those platforms.

• 10:44Alex Kantrowitz?

  Okay, so let's dig in a little bit into how your team rates. You know, it's kind of interesting. You brought up that they look at sentences and they rate based off of factual integrity. And there's someone from the right, left and the center there. But what is the actual process? Like, how are you really able to tell if a story is is like right, left or center bias? For instance, like the New York Times there skews a little left and the Wall Street Journal skews a little right in your chart. They're both like within the like the respect the fact, obviously. I guess maybe not obviously, people might have a bone to pick on that in the respect the fact category. So how do you like rate each story?

• 11:24Speaker 01

  Well, the news, overall news source ratings, like you mentioned, New York Times, Wall Street Journal, being a little to the left and to the, to the right respectively. Their ratings are based on a representative sample of content. And for each one of them, we've rated hundreds of articles manually over the last several years. And so for each article we rate, it's a panel of, of three folks. So, I mean, this is, it's as in, as work intensive as it sounds. All day, every day, in shifts on Zoom, you have panels of three analysts at a time, just going through article after article after article, episode after episode. And they're looking at certain factors for reliability and certain factors for bias.

• 12:08Speaker 01

  And it's, it's akin to like a grading rubric for, for a paper, like a, you know, AP test or something. Or if you're like, you're trying to grade something that's inherently a bit subjective, like gymnastics or figure skating, we're looking for certain things. And once folks are trained in our content analysis methodology, they can systematically look for those things and arrive at pretty consistent ratings.

• 12:35Alex Kantrowitz?

  What are you looking for?

• 12:37Speaker 01

  So the sub factors for reliability are, you know, one, the, the headlines and, and images, like how those, how well those match the story. Two, expression is a really big one. So how it's expressed as fact, as analysis or opinion or worse. And then the, the main one that think, the, the one that folks think of immediately is veracity, like likelihood of veracity. And this is the one that trips people up philosophically because they're like, well, you know, how do you tell what's true? Look, there are, there are things that are truer than others, right? And there are things that you can be 99% sure, there are things that you can be like 60% sure.

• 13:21Speaker 01

  So we're looking for the likelihood of veracity of underlying claims to the best of our ability by looking at other primary sources and other reporting on, on the internet. Like, we can do lateral reading, that's like the primary way that folks who teach media literacy.

• 13:39Speaker 01

  Teach people how to evaluate evaluate whether something's true or not. And for bias, we're looking at different sub factors, too. Like there's the language that people use to characterize their issues or opponents. Right? Like you can use, you can use adjectives to describe. You know, some politician is like senile or decrepit or cunning. Right? All those matter. There's the terminology that you use for the positions like, you know, illegal aliens versus undocumented persons. Right? We're looking for the advocacy of political positions. Some have it, some don't. We're looking at the comparison between your different articles about the same topic.

• 14:24Speaker 01

  So each one of those sub factors is one that we can use to triangulate and measure like an actual score for bias in a rubric.

• 14:32Alex Kantrowitz?

  And then you apply those ratings to different tranches of media that an ad buyer can go in and basically be like, all right, like, give me the ad Fontes group that is, you know, let me, I'm going to just pull up your chart.

• 14:46Speaker 01

  Yeah.

• 14:47Alex Kantrowitz?

  That one that, you know, might be middle, skews left and skews right, but isn't hyper partisan right or left. And maybe within like the realm of truth of, you know, opinion or wide varieties and reliability. Or, or just like, you know, most analysis or mix of is a mix of fact, uh, reporting and analysis or fact based reporting. So is that they can sort of pick like the buckets that they want and then go to an advertising technology platform and say, give me these.

• 15:15Speaker 01

  Yeah, exactly. I mean, generally, uh, for, for most brands that want to advertise in news or, you know, do advertise in news. You know, they want to stay towards like the, the top middle, not, you know, exactly right in the middle, like centrists. Like I said, there's nothing wrong with like left and right bias. And there is a certain, you know, there's certain analysis and opinion content that's very palatable to advertisers and has great audiences. So yeah, we basically allow advertisers to select and get back into news based on, you know, being on the higher reliability side and the minimally biased side.

• 15:55Alex Kantrowitz?

  Interesting. So talk a little bit about how long the company has existed, how long this product has existed, and what tangible results have you shown so far from your efforts?

• 16:06Speaker 01

  Yeah, I mean, we have been around as a company for a little over five years. We're a public benefit corporation. So, you know, we're a for-profit company with a stated public mission. You know, this all started because we wanted to help folks navigate the news landscape. And that's all the stakeholders in news media, whether it's your individual consumers, you're just figuring out for themselves what's reliable and biased, whether it's educators, teaching media literacy. But advertisers, publishers, and social media companies all have such a huge role to play in this. And our data has really come into its own in the last two, two and a half years, or that's the time period during which we've

• 16:51Speaker 01

  had dozens of analysts. We have now rated nearly 10,000 different individual news sources. So it's a lot of, it's a lot of data out there, right? And since it's now available in DSPs and is selectable, it's

• 17:11Alex Kantrowitz?

  Which is basically, for those listening, a DSP is a place you buy online ads.

• 17:15Speaker 01

  Yes, I'm sorry. I just jumped right into the ad tech, you know, the terminology.

• 17:20Alex Kantrowitz?

  Happens all the time.

• 17:20Speaker 01

  I know. It's selectable, you know, in the places where you can buy media. So we're really, we're really optimistic. We're really hopeful with what we've seen so far with advertisers, you know, major advertisers in all the different verticals, using our data to get back into news responsibly. Because that's, it's been encouraging, right? Our message has been like well received. Like, people know that they should advertise in good journalism and avoid the misleading, polarizing stuff. It's just that they haven't had good ways to do it before.

• 17:58Alex Kantrowitz?

  So the advertisers aren't coming to you and saying, we want the hyper partisan, not true stuff. Like it's mostly the folk, mostly advertisers saying, we want some, you know, reasonably biased publications that respect the facts.

• 18:11Speaker 01

  Yeah, exactly. Like, no one like wants to advertise on the hyper partisan, like misleading content on purpose. Like, I really haven't had any advertisers that are like, yes, our audience is there. Let's go, let's go after that. Like, you know, advertisers have like this.

• 18:27Alex Kantrowitz?

  You could, if you're MyPillow, you might want to use this stuff. But anyway, I digress.

• 18:32Speaker 01

  People do bring that up. And sure, they're like, advertisers, for the most part, try to block this stuff. Like, there's enough energy around, enough like self-awareness to know that they shouldn't advertise on misleading, extremely polarizing content. A lot of it ends up there on accident, right? It's, the ad tech ecosystem is opaque, sometimes not that easy to control exactly where your advertisers run, especially on the internet.

• 19:05Speaker 01

  So yeah, no one's come to us saying like, yes, we want to target the most extremes.

• 19:11Alex Kantrowitz?

  Do you have a sense as to how much money you think you've brought back into news? How much total spend has gone through with your data? And can you name some names of advertisers that are using this stuff?

• 19:22Speaker 01

  You know, it's hard to say, you know, dollar wise, but we do have, we just launched this billion dollar challenge to get advertisers back to news in this 20, especially in this 2024 election cycle. So we're still pretty early. That's a challenge that we issued just within the last month or so. And it's funny that you ask about particular brands. I'm not going to mention any names without, you know, prior express permission, because when it comes to things like, you know, political, how biased stuff is, and you know, how, how ad brands make decisions about where they advertise, as you can imagine, some are like a little sensitive. So, you know, hopefully within the next few weeks,

• 20:07Speaker 01

  for our billion dollar challenge, we'll have some brands that we can talk about that will endorse. Like, that's what we're actually looking for. We're looking for brands to step up to the plate and say, yes, like, this is something that not only we do in practice, but we believe in and want to put our name on it. So it's a You know, I want to be, you know, fully like frank about the challenge that this is, right?

• 20:36Alex Kantrowitz?

  Right.

• 20:36Speaker 01

  So many big brands are afraid of news and of putting themselves out there and saying, like, yes, we're going to draw this line in the sand about reliability and bias. That's why it's, uh, that's why newsrooms continue to suffer because brands have been so hesitant, uh, especially in recent years.

• 20:57Alex Kantrowitz?

  So then why, I mean, I guess like from my standpoint as a journalist, trying to see how real this is, like, you know, what, what should convince us that this is something that's going to actually take off and lead to a spend that helps restore, you know, it helps invigorate publications that respect the truth and aren't skewed hyper partisan, helps reinvigorate them financially and that this thing can actually accomplish its mission versus something that's, you know, a cool idea in principle, but might just never take off because of these hesitancies.

• 21:29Speaker 01

  We can't, we don't know that for sure. That's the thing. I mean, that's the, uh, that's the unknown part, you know, we're a startup where, uh, we have not, we have not yet proven that we can successfully bring back so many brands to advertising and news that it's going to make a difference in, uh, the success and monetization of good, um, of good publishers. And that's a scary place to be in, I think, for our democracy, right?

• 22:03Speaker 01

  Publishers have been struggling with this for a while. I mean, we talked to major publisher, like some of the, the biggest names out there who during 2020, 2022, um, coverage of COVID, coverage of like the, the, um, the protests after George Floyd's murder, right? The negativity in news, like everything in the news was negative, right? And publishers would say to me, you know, our, um, our readership went up, up, up, up, up. Our monetization just went down, down, down because brands were just like, oh no, we don't want to be around COVID coverage. We don't want to be about, around coverage about Black Lives Matter. We don't want to be around, around this stuff.

• 22:46Speaker 01

  And so I don't have a guarantee for you, for anybody, that my company, as much as people say that they love this idea, as they say, like, oh, yes, definitely want to support journalism and brands should support democracy. No guarantees, right? So that's why we're issuing this called action, this billion dollar challenge. It's going to take some changes. Like, it's the trend has not been good for publishers. The responsibility of brands, it hasn't been, they have not been stepping up to the plate, right? So I'm an optimist, and I think it can

• 23:29Speaker 01

  work, but we need people to jump in the boat with us.

• 23:33Alex Kantrowitz?

  Vanessa Otero is here with us. She's the CEO of Ad Fontis Media. We'll be back in the second half with some more questions about the business model, and maybe what we're going to see heading into the 2024 election. Back right after this. This episode is brought to you by App Point. Everyone's racing to roll out AI right now. Copilots, chatbots, agents doing real work. But here's the part nobody loves talking about. All that AI runs on your data, and most teams have no single way to see it, secure it, and prove it's under control. That's exactly what App Point does. For 25 years, they've been the trusted layer beneath the world's most demanding data, now extended across your entire

• 24:13Alex Kantrowitz?

  AI estate. Your data, your cloud, and the agents acting on your behalf. It's how more than 28,000 organizations deploy AI with confidence. So innovation scales without scaling risk. App Point, the unifying trust layer for AI. Head to apppoint.com to see how enterprises deploy AI with confidence. Learn more at avpt.co/bigtechnologypodcast. That's avpt.co/bigtechnologypodcast.

• 24:46Alex Kantrowitz?

  Violent threats against executives are growing at an alarming rate. When a company experiences backlash, executives are the first people blamed. When their address and other personal details are sitting online, this backlash can lead right to their front door. The team at Ironwell knows this better than anyone. They've protected some of the most targeted executives and individuals on the planet for almost two decades. As a wide-glove enterprise security service, they protect your people with continuous personal data removal, proactive prevention tools, and human-led emergency support. So when someone goes looking for your executives, Ironwell ensures they hit a dead end. So here's what to do. Go to ironwell.com/bigtechnology, fill in the quick form, and request your free risk assessment. Their team will show you just how exposed your executives are

• 25:31Alex Kantrowitz?

  and how to lock it down before a threat reaches their front door. That's ironwell.com/bigtechnology. Learn how to stop online threats before they become real-world attacks. The link's in the show notes.

• 25:51Alex Kantrowitz?

  And we're back here on Big Technology Podcast with Vanessa Otero, the CEO of Ad Fontis Media. Vanessa, you've also, you've written in depth about what you're planning to do here. So a couple more questions for you about the, what happens? The impact, you know, should this succeed? So first of all, you write, "As you veer off to the left or the right, quality and trust drop precipitously." So I'm curious, when you write something like that, like, how do you answer like the both sides argument that people might say like, "Well, like, you know, okay, like, you know, you might have some of this stuff on the left and some of this on the right, but the magnitude isn't the same." So you're putting it in the same sentence. Some people might not be fans of it. I'm curious what, I mean, I personally think that, like, yes,

• 26:34Alex Kantrowitz?

  there are problems on both sides of the political spectrum here. But how do you answer that argument?

• 26:39Speaker 01

  Well, just because there are two sides to an issue doesn't mean that they're equal and opposite. Like, if you think about a case in court, right? There's always two sides. And what we're trying to do is, you know, be, be, you know, judges of the content. And a judge, you're not asking a judge to be neutral. You're asking a judge to be fair, right? What I mean by that is, at the end of the day, the judge doesn't say like, oh, well, you both have a point. You know, good luck. You ask the judge to like, make a call, you know, which, which side has a better argument? And the, you know, side with a better argument in a particular, on a particular issue, tends to be the one that has

• 27:21Speaker 01

  better facts, more facts tied to their analysis and, and their conclusions. You're right that folks, you know, folks on the left and the right will, you know, point to the other side and say, well, they're worse because of X, Y, and Z. And the nature of what's like far left and what's far right on our chart are, are different, right? You'll notice, if you look really closely, especially at our interactive media bias chart and at our, our web, our web chart versus our podcast chart, there is a, there are more right-leaning websites.

• 28:00Speaker 01

  There's just literally a larger number, objectively, of sites that have misleading, inaccurate information. It's just like a bigger ecosystem over there. So folks on the right will point to that and say, well, that just means that your chart is biased because there's more stuff on the right. I mean, I don't think that's what it does mean.

• 28:26Speaker 01

  The, it's not that there's not misleading content on the left. And there's just less, there's a lot less website content that's specifically tailored towards left-leaning, misleading, inaccurate information. But you go to podcasts, though, there's a, there's a nice, like, little universe that's long-form. You see this in video, too, as you can find a lot of YouTube channels that are left-leaning, that are conspiratorial, right? So it's just not the exact same content universe on each side. So, you know.

• 29:02Speaker 01

  It just be just because two sides exist doesn't mean they're doing exactly the same thing. On the right,

• 29:09Alex Kantrowitz?

  Right.

• 29:09Speaker 01

  we do see, um, while there will, while there is more, you know, misleading and inaccurate content, you'll rarely see like bad words, for example, that's something that on the right folks tend to refrain from. However, on the left, like, um, really partisan, you know, left-leaning opinion propaganda stuff, uh, tons of curse words, tons of, um, you tons of really vile personal insults. So that's not exactly the same thing. That's those are just the things that move things towards the bottom of the chart on each side.

• 29:44Alex Kantrowitz?

  Another argument that folks might make is, well, what happens if this, uh, company ends up skewing, you know, some of these big brand dollars like to one political side or the other. So is your goal, uh, to proportionally direct ad dollars to like, sort of, uh, reasonably biased right and left sites? Or like, what happens, for instance, if a lot of the money that ends up coming in, you know, using your data ends up going to left-leaning sites or, or right-leaning sites? Is that still a win for you because they're going to more reasonable news? Or is that something that you want to stay away from?

• 30:20Speaker 01

  We want to direct, uh, ad dollars towards high more highly reliable sites. And there are highly reliable sites on, on the left and the right. So, yeah, and I don't want to, you know, duck this question. I, I talked about how the, like, the right-leaning sites, uh, how there's more like misleading, um, in inaccurate ones, like below, like a low threshold on our chart. Well, one, uh, so folks on the left look at that and they're like, ha, ha, see, I was correct. Um, but the chart also shows data that, um, it vindicates folks on the right. A lot of folks on the right say that the mainstream media, like some of the biggest publications, skew left. And if you look at the concentration of some

• 31:04Speaker 01

  of the biggest news properties in, in our country, uh, they do skew left. They're not sitting like right there on the middle of, of the chart. And it's

• 31:17Speaker 01

  There's not an equal distribution above this, you know, the high reliability threshold. And if you think about like how our country is and how content is and how there's just tens of thousands of news source out there, it's almost an impossible expectation to think that there would be equal and opposite news sites across every possible spot in the chart, right? There are so many different factors at play. So ultimately, we just want folks to support, brands to support the most reliable content, whether it's left center or right leaning. And I say center because center is a bias too.

• 31:59Alex Kantrowitz?

  Vanessa, is there a risk that this just ends up like directing money to the mainstream, mainstream sites versus like the upstarts? I mean, because it does seem like, all right, like part of like a proxy of what we've been talking about is like, you know, these are, are these establishment sites or, you know, talk, so talk about your rebuttal of that, what people would say on that front.

• 32:18Speaker 01

  So some, somebody can score high reliability on our chart if they're like brand new, you know, they do not have to be around for a long time. And there are a bunch of newer sites, just in the last, in the last few years, like, like the messenger is a new site, you know, that popped on, on the scene and we rated it as a high, has a high reliability score. But it doesn't have to be just like a big site. That happens to be one that's like well-funded and it has some like media industry veterans associated with us, but being new doesn't preclude them from being high reliability. But also being small doesn't preclude you from being high reliability.

• 32:57Speaker 01

  So podcasts, like, yeah, we, we scour like the, the media landscape for, for new substacks. Like it, it can be just a one-person shop, but if that content is about the content. So we actually are helping advertisers discover not just like the good stuff they know, you know, it's, you know, do I need to tell people like, oh yeah, the AP and Wall Street Journal and like New York Times, like those are good, people know those ones, right? We're talking about the ones that people don't know, the new stuff, the, the small stuff, the local stuff. You know, there's so much good local content. There are so many good trade publications that absolutely deserve this kind of, this spend, these ad dollars.

• 33:41Speaker 01

  And we have like thousands and thousands of them that we've identified.

• 33:46Alex Kantrowitz?

  Okay, can this extend to social media at some point? I mean, is this something that like a buyer who wants to buy? Like we, let's go full circle. We talked about Facebook. We talked, you know, we talked a little bit about X. Can a media buyer like there's been so much discussion right now in terms of like, you know, Elon Musk and his, you know, anger towards advertisers for pulling out of Twitter completely. Like, could this be something that an advertiser decides, hey, I'd like to go to Twitter, but want to buy mostly like the discussion that falls into the, you know, less extreme areas of conversation? Or is that too complicated to work through?

• 34:22Speaker 01

  It's actually not. So we actually have data on X, X Twitter handles and publications, because everything that we've rated on our, on our chart, where they have like Twitter handles and other social media handle like YouTube, we, we match it up and tag it because the content that somebody produces outside of social media is often very similar, if not the same, as the content that they put on social media. And, you know, X has adjacency controls. You know, they have a tool, for those who aren't familiar, like for advertisers,

• 35:02Speaker 01

  they can say, the advertisers on X can say, I do not want my ads showing up next to, you know, these particular handles or these particular keywords. We, we have data that advertisers could use to make those decisions on platforms like that. And ideally, we'd like to see that. And we have that kind of data for X. We have it for YouTube. We even have it for threads, even though I don't have a ad product yet. But yes, the-

• 35:32Alex Kantrowitz?

  Has any advertiser used that yet?

• 35:34Speaker 01

  Not yet. So it's a new, it's a new offering. So, you know, we're hopeful that we've, we're hopeful that folks will because social, like as much as, you know, social media is a really fraught place with, you know, lots and lots of real valid concerns about the content that's on it broadly. Advertisers sort of can't help the wanting to reach their audiences that are there. And, you know, different people see different stuff across those platforms.

• 36:07Alex Kantrowitz?

  Let's end with a, you know, a small discussion on AI.

• 36:09Speaker 01

  Yeah.

• 36:09Alex Kantrowitz?

  You have dozens of people under your employ that are rating these stories. You mentioned it's like a very painstaking process. Can artificial intelligence be something that we could reliably use to rate stories? I mean, AI comes with its own bias, so it's kind of interesting. How do you see AI factoring into the future of your, of your work?

• 36:27Speaker 01

  Well, we've been working on this for, this problem for a really long time. And to be honest, a lot of folks have, for a long time, tried to, like this holy grail is like, oh, what if we could just identify misinformation with AI? Wouldn't that be so great? Like, we would have none anymore. And the reason that that hasn't happened is because it's very difficult for a machine to tell like the nuance of how true something is and how left and right it is to degrees. So, you know, that's what we started with manual ratings, but you know, over the years, we've developed it to, it's almost over 70,000 individual pieces of content that we've hand labeled.

• 37:12Speaker 01

  I'm a patent attorney by background. I, you know, did software patents in my career. And I knew that, you know, manually labeled training data is what you need in order to create AI for, for pretty much anything. So we have this, the largest in the world, we believe, set of labeled data for this. So we've developed our own AI models. So we can actually now score articles at scale for reliability and bias. And it's quite accurate as compared to our own human ratings. I say this is something we've been working towards for a long time because, you know, folks look at the painstaking work that we've done to rate articles and episodes.

• 37:57Speaker 01

  And they're like, wow, that's, that's great. Like, can we do this at scale? Because in advertising and social media and content on the internet in general, there's just so much of it. So humans aren't scalable enough. AI is not accurate enough. So for us, it, you know, I'm really, I'm pleased that we've been able to get to this point where we rate the top news articles with humans every day and we rate the rest of the news landscape with AI. So we're rating now tens of thousands of articles per day to help advertisers like reach that scale that they want.

• 38:34Alex Kantrowitz?

  What, okay, I just got to ask you, what, there's, there's, there's just been this, you know, I already said we're going to sign off, but you know, it just, it just popped in my head like that. What do you think about the whole argument, like the free speech argument? And does this sort of impede on the free speech argument?

• 38:49Speaker 01

  I, it wouldn't surprise you to note that I've thought a lot about this. So, um, especially like when it comes to content moderation and, you know, section 230. And, you know, for, for those not familiar with, uh, you know, section 230, I mean, it's, it's a law that basically allows social media companies to escape liability for either moderating content or not moderating content, like let social media companies just sort of off, off the hook. And the, it seems that the only real workable way to do, do, do anything about misleading

• 39:30Speaker 01

  or extremely polarizing content on the internet while still respecting not only the laws of free speech in this country, but the ethic we have around free speech in this country. The only way to respect that is to label content and provide users more of a choice. So, you know, people, no one in this country, I mean, this is a broadly shared left and right sentiment. People don't want information to be repressed. People don't want information to be censored. They want to be able to have choice about information, even if it's abhorrent or even if it's false. But the reality of like the, this very overwhelming information landscape is that people need more

• 40:13Speaker 01

  information about the information that they're going to consume. So you're, it's a, I view labeling stuff as a solution, like it's like a content rating for movies. Like this is G, PG, PG 13, R, like you just have a general idea when you're walking into it, like what you're getting into. You can have the choice whether to, to do it or not. So like a label of like, you know, left and right, here's how far this is fact analysis opinion, or it has some other problems. Here's some more information about it and you can make that choice about it. To me, that's a version of more speech being the solution to your free speech problems.

• 40:55Alex Kantrowitz?

  Right. Okay. Really, the last question. Any 2024 campaigns coming to you to say, hey, I want to advertise to like this segment of the population? And if not, do you expect them to come through?

• 41:05Speaker 01

  Yeah, we have had some agencies for political advertisements express interest in our in our ratings. And we find that really fascinating. So just to be clear, like, the terms of our data are really explicit that you cannot use our data to target like most extreme or misleading content, right? It's really fascinating to be able to use our our information about content and how left and right it is, because we believe that there's a very high correlation between like the content people read and their political views, right? Usually, center people read center stuff, and center-left people read

• 41:49Speaker 01

  center-left stuff, and center-right people read center-right stuff, and on and on, right? And it's hard for political advertisers to target people to that level of voters to that level of granularity. And in elections that are divided that are decided by just tiny vote counts, tiny percentages, it's really important to reach persuadable folks. So our data is quite useful for identifying who would be persuadable. And hence, it's folks that read more reasonable content.

• 42:23Alex Kantrowitz?

  Yeah, like if I'm in a Republican primary campaign, I'm working on that staff, I'm going to you and saying, get me the, you know, people that are reading, you know, the based in fact, center-right publications, and give me them in Iowa. And, you know, then you're really talking about a group that can have weight politically.

• 42:45Speaker 01

  Yeah, absolutely. I mean, the folks that are like reading very, very strong left and right, they tend to not be persuadable. It's just the nature of our, you know, very polarized society right now. But there, what gives me hope is there are a lot of reasonable folks reading high quality, high reliability news. And yeah, we want to elevate the folks who are putting out that kind of work, that kind of journalism.

• 43:15Alex Kantrowitz?

  Vanessa Otero, thanks so much for joining.

• 43:17Speaker 01

  My pleasure. Thank you, Alex.

• 43:19Alex Kantrowitz?

  Thanks, everybody, for listening. Thank you, Nate Guattini, for handling the audio. Thank you, LinkedIn, for having me as part of your podcast network. We have a great set of shows coming up for you over the next couple of weeks. We do hope you stay tuned, and we'll see you next time on Big Technology Podcast.

• 43:41Speaker 02

  Now, more than ever, we need someone on our side. But Rob Whitman? Whitman's votes drove up costs, cut Medicaid, and risked 350,000 Virginians' health care to pay for tax breaks for himself and his donors. He's not on our side. Shannon Taylor will fight corruption, cut taxes, and bring down costs. This November, vote for Shannon Taylor. She's for us. Paid for by HMP, the House Majority PAC dot com. Not authorized by any candidate or candidate's committee.

• 44:13Speaker 03

  Every mile matters. Whether you're driving your dream car or your daily commuter, the Mobile One team knows that. It's why we created Mobile One Advanced Clean Motor Oil. It cleans your engine in one oil change and helps extend engine life, delivering performance, protection, and cleanliness for vehicles of all ages, all to help keep your engine running like new. Whatever you're driving, visit mobileone.com/advancedclean to learn more.