Episode description:

One reason that falsehoods flourish online is that major advertisers fund them—but usually unwittingly. The opaque nature of automated online ad delivery means that advertisers don’t actually know where most of their digital ads appear. On a high-quality news site? Maybe. On a trashy clickbait farm? The ad-tech doesn’t care.

Today’s In Reality guests argue that quality journalism needs a more transparent market to prosper, that’s what they aim to provide. Vanessa Otero is an IP attorney turned entrepreneur, the founder and CEO of Ad Fontes Media. In Latin, the name means “To the Source.” Vanessa is joined by her CSO Lou Paskalis, who among other roles was a senior VP of media investment at Bank of America. For the two, the work of steering ad dollars back to quality starts with a unique media bias chart, which ranks thousands of news sites, television, podcasts, and newsletters by quality of journalism and degree of political bias.

Ad Fontes Media bias chart: https://adfontesmedia.com/

This episode was produced by Tom Platts 

 

Episode Transcript

 

Eric Schurenberg (00:01.981)
Vanessa and Lou, welcome to In Reality.

Vanessa Otero (00:05.236)
Thank you so much for having me.

Lou Paskalis (00:07.522)
Thanks, Eric.

Eric Schurenberg (00:08.741)
It is great to have you both here. Now, let’s set some groundwork. I’m not sure that everybody listening to the podcast completely understands the tech ecosystem. I know it’s pretty complicated, and I lived through it at Inc. and Fast Company. I have this suspicion that people’s general understanding of how ad markets work has a much change from watching Mad Men. So if you see an ad for Chevrolet on Fox News or YouTube or… or CNN online, you might assume that advertisement is there because the brand intended it to be there. And while that’s true in the shrinking categories of print and linear TV and digital advertising where most of the spending is, the mechanism behind that ads appearance are pretty opaque, driven by technology and human input doesn’t do much more than set the parameters. So…

Talk us through the process. What are we, you know, what is actually working behind the scenes in the ad market today?

Lou Paskalis (01:14.09)
Yes, I’ll take that one, Eric, having been a buyer for three and a half decades. The change is dramatic, as you suggest. What used to be a transaction between humans has now, particularly in digital, become an automated bidding marketplace where the entire sequence of events from inventory becoming available to bids being placed to a bid being selected.

to an ad being trafficked takes place in 30 milliseconds. You cannot blink your eyes as fast as that. And this has allowed marketers to tap into greater audiences, to leverage more real-time insights, and to actually create a price value equation that, on the surface, is favorable to them. Unfortunately, there are also a lot of other things that have gone from being very transparent to very opaque.

I don’t know where all my ads are running, even though I’ve introduced many safeguards to ensure that they only run in the places I want them to. I really don’t know what the final price of an individual ad is, because I’m looking at a blend of pricing, which may include some very low cost environments or inventory that may be unsavory or not the kind of places that I want my ads to run.

And I think for me as a buyer, the thing that’s most vexing is that in the old days, I could call you up as publisher of, you know, Fast Company and say, hey, Eric, I don’t like where that ad ran. I think you owe me a make good. And then you’d take me to lunch, ploy me with wine and convince me I was wrong. But it was a nice wholesome process. But we’re now talking about billions of transactions in a millisecond. And so that’s no longer possible to pragmatize.

Eric Schurenberg (03:08.685)
Okay, that is great. And just to clear up some industry jargon, inventory is basically a page appearing, being served to a reader, a user of a website. And then the bidding that takes place is that the ad technology alerts a machinery that an inventory has just come and then automatically there’s bidding for the eyeballs, the attention of that user.

So, okay, I understand that now that this is instantaneous, that it’s no longer that kind of comfortable transaction over lunch that you just described, Lou. But still people are making decisions, advertisers, brands are making decisions based on what they think will help their advertising be most effective. So, I don’t see any… connection here necessarily to the spread of falsehood or misinformation online. Where does that connection come in and why has that brought Ad Fontis into business? Vanessa, would you take that one?

Vanessa Otero (04:18.096)
Yeah, well, what we do at AdFontes fundamentally is we rate news and information sources for reliability and left to right political bias. And when we say news and information sources, that means everything from the best journalism to fake news to blogs, to opinion, websites, podcasts, television shows, wherever people get their information sources.

Vanessa Otero (04:47.88)
in this work. I did not come from the ad industry. I was a patent lawyer prior to starting AdFontes. I have a real passion for bringing people together. That’s why I got into this work, to identify misleading content, inaccurate content, and especially highly polarizing content. But it wasn’t long before I started doing this work that folks in the ad industry started taking notice and said, hey, you know what? There is a brand safety.

There’s a brand safety play here, aspects to your business. And I was like, well great, what’s brand safety? And lose marks because this is a really hot topic in the ad industry. It’s only been around for about 10, 12 years or so, but it’s the concept that you don’t want your ad showing up next to content that is unsavory in various formats. And that makes sense.

Trace this back to, you know, back in 2016-17, there were beheadings on YouTube and advertisers were running next to them and that was horrifying, right? So a lot of technology has arisen to make sure your ad doesn’t show up next to violent content like that, or sexually explicit content, or racial slurs, like other hate speech, and that makes sense, especially on social media where there’s just so much of it.

in this user generated content. So the work, but it’s sort of expanded this brand safety technology to include everything that might be somewhat unpleasant. And unfortunately that includes news according to some technologies. So here’s how it came into play with my company. You know, we went around to brands because folks let us know, hey, you know,

advertisers don’t want to show up next to misleading content, next to polarizing content. And we said, great, let’s go to brands and they can use our data for that. And that’s certainly a problem. If you go to some of these websites on the internet, you can certainly see some big brand ads. But there’s less and less these days. And we went to, as we started going to brands, we said, hey, we can help you solve this problem. And the thing we were surprised to learn was they would say,

Vanessa Otero (07:12.388)
oh great, but that’s not a problem for us, because we just don’t advertise next to news anymore. And I said, what do you mean you don’t advertise next to news? I mean, news content is about 20 to 25% of all content on TV or the internet or on audio. You just don’t advertise there? And he said, yeah, I mean, it’s risky, and we do all this brand safety blocking. So it’s an alarming problem. Now you’ve got the, you know,

the worst of all worlds, where you have the cheap sites that can be run by like one, two, three, four people, create some polarizing, rage-generating content. It doesn’t take much of them to skim some money off of this complicated ad tech ecosystem, get a few digital ad dollars flowing their way. Brands don’t even know, and they survive. However, the expensive journalism, like the high-quality stuff that requires staff, editorial,

You know, newsrooms, big brands have been pulling away from that whole category. They’re really suffering. So that’s what we’re here to fix.

Eric Schurenberg (08:19.173)
Okay, all right, so let me make sure I’m following the story here just for the sake of clarity. In the ad tech ecosystem where you are a brand you may not know for sure where your ads are showing up, the technology is effective enough but also kind of a blunt instrument. So you can say, I don’t wanna be next to, you know, violent content or hate speech.

But the simplest way to make sure that never happens is to say, I just don’t want to be next to news. And then that creates a lot of perverse incentives and essentially defunds some of the most important, high-quality news organizations out there that are really important for keeping an informed citizenry plugged in. Fair enough?

Vanessa Otero (09:14.768)
Exactly. I mean, one of the outside of categorically blocking news, one of the most popular tools is keyword blocking. So you can, it’s just a dumb way to do it. You can block the words, you know, killing or shooting or Trump or LGBT or Black Lives Matter. And all of a sudden that just wipes out tons and tons of coverage. You know, articles can say, you know, this person is absolutely killing it in their industry. Well, that’ll get keyword blocked. Or shooting can be about a basketball game and that will get blocked. So.

Eric Schurenberg (09:49.902)
Mm-hmm.

Eric Schurenberg (09:55.393)
Mm-hmm. So, yes, it is a very blunt instrument. Now, you’re trying to steer advertisers back to mainstream reliable news by rating news brands for reliability and for political bias, as you mentioned earlier. Tell us about how that process works.

Vanessa Otero (10:15.476)
Well, we really believe in a human plus AI blend to rate news, because AI is not accurate enough to capture nuance, but humans just aren’t scalable enough to address all the content that’s out there. So for years, we’ve been working towards the solution that we have now, and it’s all based on, you know, fundamentally the work day in and day out of humans rating content.

Ad Fontes means to ‘the source’ in Latin, and it’s about rating the content for reliability. It’s not about like opinion polling of consumers, because if you ask opinion polls like, do you trust Fox News or do you trust MSNBC? Well, the result of that is going to tell you a lot more about the person you’re talking to than the content. To get an objective as possible, look at how truly reliable or how truly biased something is, you have to be looking at something repeatable, which is the content on the page. So the headlines and the images that are used, the very sentences, like how they’re expressed, how true they are, like the underlying claims, implicit claims. When you’re looking at bias, you’re looking at the language that’s used to characterize issues and opponents. You’re looking at whether it advocates for political positions, or how it compares to other articles about the same topic.

So we’ve had over 120 human analysts as employees of AdFontes Media over the years participate in our ratings. We currently have about 60 on our staff right now and they are politically balanced, left, right and centre, which is a hugely important part of our work to remain balanced, to get as many perspectives as possible. These analysts are trained.

in our content analysis methodology. And they spend all day, every day in shifts on Zoom, just like us here, right? Three people, one left-hand, one right, one centre. And they’re looking at lists of articles, you know, podcasts, TV shows, and they’re rating it according to a rubric, entering it in our software. And to date, we’ve done about 70,000 individual pieces of content like that over the last three or so years. So we’ve used that manual training data to train AI to rate news articles like humans do. And because it’s based on this training data, it’s really impressed with how highly accurate it is. It comes close to the standard deviation of our actual human analysts. And so we use that combination of humans plus AI to rate the individual content, which rolls up to the individual news publication itself. So it doesn’t matter if it’s a long-standing publication that’s been around for a hundred years in print. It doesn’t matter if it’s a brand-new substack from a reliable author. We rate the content the same if it’s reliable. It gets high scores.

Eric Schurenberg (13:22.381)
There are other organizations out there that rate media bias. NewsGuard is one that comes to mind, a for-profit company. Check My Ads is sort of in the same business as AdFontes, in connecting advertising to the content and then talking to the brands. How would you compare your news rating system with NewsGuard or All Sides, another organization that that attempts to rate media bias.

Lou Paskalis (13:52.414)
Yeah, you know, Eric, I, you know, I kind of like to share the epiphany that I had having been a NewsGuard client for many years at Bank of America and not realizing the subtle nuances that really matter in how news is rated. NewsGuard, which at the time was really, you know, the most well known of the news rating services.

Eric Schurenberg (14:04.222)
Mm-hmm.

Lou Paskalis (14:21.194)
does an annual survey, looks at a handful of articles, and makes a platform level determination. Nothing wrong with that, right? But at the speed of culture today, where something is in a moment and out in the next moment, that really kind of does not age as well as the methodology that Ed Fontes uses. And I didn’t realize that until I joined Vanessa as her chief strategy officer. Our platform ratings are an amalgam of our most recent

article ratings on that platform. So we’re constantly looking at video stories, articles, podcasts that come out on that platform. And each time we rate one, that platform moves in whatever direction that most recent rating will take it, all very nicely calibrated, not to over-emphasize. I think the most dramatic move that we’ve seen in the last half a year is when Tucker Carlson left Fox News, the entire platform shifted a little bit up, which is to say a little bit more reliable, and a little bit to the left, which is to say a little bit less biased to the right. That can only happen if you’re in a continuously ongoing monitoring process. And we haven’t really seen anybody who does it the way that we do it. And it really allows us to be very relevant all the time. And I think it’s a testament to Vanessa’s original vision when she set up the business.

Eric Schurenberg (15:49.101)
Okay, thanks. Let’s dig a little deeper into the standards that you use for each of the axes, the reliability axis and the political bias axis. Let’s talk about political bias. There are, you know, left, right, as you said. Tell me more, if you can, about how you calibrate that, Vanessa?

Vanessa Otero (16:12.692)
So that’s a great question. And it’s one of the fundamental questions that we get asked when people spend any time at all with our media bias chart. Because the reason it’s popular, I have studied this concept, this phenomenon, why this is popular over the years. Because it sort of caught me by surprise that it was. I originally created it. It was a hobby. It went viral. I thought, why?

And it’s because it’s a two-dimensional visual image. That conveys a lot of information in a short period of time. And it’s just this balance. Like we’re used to seeing infographics. And it’s this balance between complexity and nuance and simplicity. Because you have to balance those things. There’s a whole field of like media literacy and political science. You could read academic papers on polarization and why our news media is how it is, but the reality is most people are not reading that. Most people sharing information on the internet aren’t getting PhDs in information literacy. So people will share something from Infowars, Alex Jones content, as well as they’ll share something from the AP. And some people have a hard time distinguishing.
So the bias axis is intuitive at first. Oh yeah, I get it, left, right. America, left and right. And people will make assumptions about what that means because left and right is really the broadest set of terms that you could use for that with political bias. We could view it as liberal, conservative, or Democrat and Republican, but there are shades of each one of those, right? So there’s progressives and neoliberals and centrist, and you know there’s libertarians, right? There’s folks, it’s a gradient. And capturing the gradient on a scale was really important to us. When we first started figuring out, alright, well what are the… The media bias chart is two things. It’s a taxonomy, it’s a system of categorization, and it’s a methodology for placing the content on that taxonomy.

So the taxonomy decisions about the left-right bias, we had to make decisions like, what is this anchored on? Because what’s left and right isn’t the same across the world. What’s left and right changes over time. It’s related to this concept called the Overton window. And if you’re going to measure media against the backdrop of what’s left and right in the U.S., you have to anchor it on something else than the media itself.

So we made decisions that the media bias chart is a US-based media bias chart based on the US left-right political spectrum. We have some international sources and we’ve done media bias charts in other countries and I can talk about how we do that. But this one that people are familiar with is a US-based one. It’s also a contemporary US left-right, which changes all the time. Because if you think back to, you know, prior to 1964 when the Civil Rights Act was passed, there were left and right positions on whether segregation should be legal, right? And the right position, like, position on the right would be segregation, you know, should be continued to be legal. The position on the left was that it should not. The center, there wasn’t really a center position. How do you really compromise on something like that, right? So the best arguments won out. We decided as a country that the segregation was not, should not continue.

And so now that’s basically like a widely held position in the U.S. It’s not like the right leaning position anymore that there should be segregation, except on maybe fringy, right? So it shifts over time. And we’re trying to capture what it is now. And yeah, sure.

Eric Schurenberg (20:26.025)
Vanessa, could I interject a question here? Just also an explanation that the Overton window is a phrase that sort of describes a range of beliefs that are considered mainstream and worthy of consideration in mainstream media. And some things that are outside the Overton window are just considered too fringe to be discussed by serious people, theoretically. But I’m, but on the, well, go ahead.

Vanessa Otero (20:51.792)
Yeah, it’s interesting. Yeah. Good. No, go ahead.

Eric Schurenberg (20:57.489)
My question for you is about, what it sounds like you’re headed towards is explaining kind of the milestones or the marker points of defining bias one way or the other. And so my question is, so right now at this moment in time, is your position is a brand’s position or a content, an article’s position on abortion or Ukraine support or any other number of issues that are dominating the news. Is that how you determine bias?

Vanessa Otero (21:37.976)
We use the sub-factors of the content to determine bias. So the categories, there’s a middle one, middle or balanced bias. There’s skews left and skews right. There’s strong left and strong right. Hyperpartisan left and right, and most extreme left and right. And how we place it within those categories, the content, is by looking at the language, the political position, and comparison, which is topic selection and biomission.

Vanessa Otero (22:08.104)
So there’s nothing wrong with being left and right. But once folks get outside these strong left and right categories, for some brands, they don’t want to associate with strong left and right ideologies. Most brands, getting into hyperpartisan right or most extreme right, you certainly don’t want to. So that’s how we make those decisions. That’s how we place the content and how brands make the decisions based on that content.

Eric Schurenberg (22:39.269)
Okay, all right. Let’s shift to the vertical axis in the chart. And by the way, we’ll put a copy of the chart in the show notes so that people can see for themselves. But right now we’ll ask people to envision the vertical axis, which is reliability. And of course, many of the brands that people are most familiar with are a mix of straight factual news, you know, X, Y, Z thing happened.

Eric Schurenberg (23:09.309)
And analysis, XYZ thing happened and therefore it means something else or just that. And then opinion, which is strictly the opinion of the columnists. How do those measure up? Do you rate each one separately for reliability or you can explain it?

Vanessa Otero (23:28.904)
We do. We rate each individual article based on whether it’s fact reporting or a mix of fact reporting and analysis or analysis or opinion or towards the bottom of the scale what’s not merely opinion but something worse than that and then things that are misleading and inaccurate. So I just describe from top to bottom on the chart fact reporting, fact reporting analysis going down, analysis below that, opinion below that selective, incomplete, propagandistic below that, and then misleading and then inaccurate. So we rate the content regardless of how the news outlet labels it. Because some news outlets, especially long-standing ones who are trying to do the right thing, label their content. However, there are new analysis sites like Reason Magazine, a libertarian publication, or Vox with a V, that they’re mostly analysis. They’re just not labeled as such. Then you have some really terrible hyper-polarizing sites that label things as opinion, and really they’re much worse than opinion. They’re dehumanizing and vilifying the other side. You know, calling them animals and aliens and not human, right?
We rate that as lower than opinion. When I talked about this area, like this area worse than opinion, there’s a big gray area in our discourse that’s not quite, it’s not easy to clearly label it as misinformation. And we actually avoid the term misinformation because it’s so co-opted and it’s so polarized itself, which I would be really precise. A lot of that stuff falls in the gray area.

Eric Schurenberg (25:22.885)
Okay, all right, good. What is very attractive to me is that your rating system takes away the incentive to produce low quality news by interrupting the flow of funding for it. You give advertisers a reason to stay away because you identify a site. And yes.

So what is attractive to me about Ad Fontes is that you take away the incentive to produce low-quality news by interrupting the flow of funding for it, to give advertisers a warning that a site is not reliable or that is hyper-partisan.

And also you’re giving advertisers a reason to come back by making them more comfortable with mainstream news. But it seems to me that brands still find plenty of outlets for their advertising without news. They can be around search, for example. They can be around social media. It feels like Ed Fanta still relies on advertisers to want to be good corporate citizens without providing necessarily an economic incentive for them to support news. How would you respond to that?

Lou Paskalis (27:09.846)
I think, you know, if you don’t mind, Eric, I’ll take that one. We’ve laid out the business case in a series of op-eds over the summer and, you know, have our own proprietary research. We’re just getting ready to release round two and at the end of the month. Essentially, the business case for news is unassailably strong and it kind of comes in three broad buckets. First of all, what you’ve got is a very attractive, the most attractive, I would argue, demographics of any vertical. You know, when I started in this industry at the dawn of time, the most valuable inventory, the most valuable space you could buy to run your ads was in the New York Times, particularly against their op-ed pages, right? We knew that. Why? Well, because the readers of those pages had the highest household income. They were the most well-educated. They were the most likely to travel abroad.

Offer pleasure they were the most you know you just on and on and on those stats haven’t changed in the thirty five years and i’ve been in the business the people who read news are being most affluent and the most influential people indian i’d states in general speaking abroad as well so if you want to sell goods particularly high-end goods those are the people who have the propensity to buy them
Because we’ve seen this tremendous decline in advertising supportive news, the statistics are damning. In a 15-year period ending with the dawn of the pandemic, there’s been an 80%, 8-0% decline in advertising investments in news, which has resulted in a 50% decline in the number of people working in newsrooms in roughly the same time period, according to the US Bureau of Labor Statistics. You now have what we call in the advertising industry as unduplicated reach. You certainly have not seen the demand for news fall off in any proportion to like the investment in news. But there are a lot of people, myself included, who are news junkies. I wake up reading news sites. If my TV’s on in the house, it’s on a news channel. I’m a constant consumer of news. And I’m not alone. A lot of people are like that.

Lou Paskalis (29:31.218)
And so whenever you want to grow your business, one great way to do it, as you know, from your background is finding unduplicated reach, people you’re not reaching with your other advertising, and bring your ads to them. And this was always the key to the success that I had in my career, whether it was at American Express or Bank of America. And then the last thing, because of the decline in demand for news, it’s now being sold at a significant discount to other verticals like lifestyle and sports, et cetera.

Good economic value of news in terms of who you reach, the amount of people you can reach that you can’t reach elsewhere, and the price that you pay to reach them has never been stronger. That’s our business case. Then we move to the items that you just said, which are we can help you find the most reliable, least biased news that comports with your values. You know, and it’s not just your values, it’s your customers’ values too. People are shopping their values more than ever before. And so where you show up matters, right? And so we can make that case. But it’s always gonna be a business case first, followed by your civic responsibility and your opportunity to appeal on a broader set of values.

Eric Schurenberg (30:47.677)
Great, thank you. Thank you, Lou, that’s very helpful. You alluded to the catastrophic decline in advertising, supporting news and the hollowing out of newsrooms around the country. It’s particularly bad with local news. And there the ramifications are dire. There are, I think, 25% of US counties have no local news. And the creation of a term news deserts has now become kind of a cliché. I noticed that most of the brands in the Ad Fontes chart are national news organizations. Do you rate local news organizations in a way that would help kind of drive the argument, Lou, that you just laid out?

Vanessa Otero (31:36.128)
Absolutely.

Lou Paskalis (31:36.535)
The short answer is yes, but Vanessa’s probably the better person to give it.

Vanessa Otero (31:42.128)
We’ve rated well over a thousand local news outlets. That includes not just the website, digital print, but the TV content as well for most major markets and some local radio as well. And it just doesn’t, if you look on our free version of our interactive media bias chart, or the infographic that floats around, those only show about 120 of the news sources we’ve rated. We’ve rated well over 3000.
We have data on another 7,000 sites across the news landscape. So yeah, we absolutely encourage, this is a real passion of mine, local news. And because of problems that plague the national news at advertisers with brand safety and brand reticence to advertise there.
Local news has three or four more additional obstacles to getting funding from brands.

Eric Schurenberg (32:45.968)
What are those obstacles?

Vanessa Otero (32:49.556)
Well, one of the main ones is this programmatic digital ecosystem for running ads alluded to earlier. In the digital transformation, remember 30 years ago when the internet started and folks were wondering how are newspapers going to adjust to this from print to digital? Wonder how that’s going to turn out. Well, we can sort of look back and see how it did turn out. And it’s kind of ugly for local news because they got their lunch eaten at many stops along the way and continue to today. So one, the big papers transitioned to digital, the Times and New York Post, and a lot of them figured out how to monetize with subscription paywalls. Then Google and Facebook for distribution really became like a primary way of driving traffic. However, these folks who made it online, they got traffic, but not the kind of revenue that they could get from the previous traffic they had gotten in their print publications. I mean, that was a big chunk of it, the revenue going over to other, like Google and Facebook, basically. But then when this programmatic ecosystem evolved, again, this is 2007, 2008, 2009, this technology that allowed for real time bidding on news sites was supposed to make it easier for publishers to get advertising from brands. Brands got used to the fact that they could now easily buy a lot of digital inventory and get what they wanted. Characterize as scale, right? They can buy millions of impressions all across the internet much faster and cheaper before. Well, not all local publishers are plugged into that system. Not all small-town journalists, you know, small-town newspapers have the technological infrastructure, especially since they’re, you know, busy reporting on the community, right? And trying to struggle elsewhere. That programmatic ecosystem created a new class of competitor for these websites as well.

A whole class of made-for-advertising websites. So there are folks who are extremely knowledgeable about this ad tech ecosystem, and they said, oh, we could just make a bunch of sites and give them all the characteristics that are appealing to advertisers, give them that scale they want, and basically compete for the traffic that local news publishers could be getting.

Vanessa Otero (35:39.484)
um, the ANA transparency study, he can speak to that more about how much these made for advertising sites are skimming away from the ecosystem in general. And the, who’s losing out on that is local journalism.

Lou Paskalis (35:54.323)
Yeah, let’s just take a moment on that, Eric, because I think it’s important for your listeners to understand that. According to Statista, which is an industry website that tracks all manner of events in advertising, 93% of digital display ads, so non-video, non-search ads, will be… transacted in the programmatic ecosystem this year, right? 7% will be transacted in sort of direct to publisher deals. And Vanessa and I recently attended a great event by the Lenfest Institute for Journalism in which we met with over 100 local publishers who overwhelmingly said, you know, I’m not in that marketplace either because of technological constraints or because they, I think, fairly accurately see it as a, race to the bottom in terms of pricing for premium inventory. But, you know, they’re sort of sitting on the sidelines of 93% of those dollars. Meanwhile, these made for advertising sites, FNS is talking about are actually now one in five impressions that advertisers are delivering in the study that we just completed at the Association of National Advertisers. About 14% of their dollars and 21% of their impressions went through made for advertising sites and how do they do this?
Yeah, so MFAs use made-for-advertising sites that Vanessa was talking about use sophisticated algorithms and increasingly AI to determine what keywords most advertisers block and then reconstruct stories that more often than not, they, let’s say, borrow from legitimate news sources but omit those keywords. And so you know, as someone with a journalistic background, the journalist really sweats over every word to make sure that it is the best reflection of the situation at hand to educate the readers about whatever the topic is. And MFAs avoid the words that might actually offend a brand’s brand suitability keyword blocking list. And so when you’re in the auction, in that programmatic auction, you’ve got this great piece of edit, let’s say from the Washington Post, which is a fulsome account of what has happened, being sold at a premium as it deserves. And then this MFA site, which is selling inventory at a discount and has no offending words in the article. That MFA site is going to win every time in a programmatic auction. And so we’re actually penalizing.

Those sites that actually have Pulitzer’s lying around their offices like paperweights and favoring other sites where the content is either borrowed, perhaps plagiarized, sanitized, and might even be written by software as opposed to humans. It’s a toxic cocktail which just really disenfranchises the fourth estate in our country and no longer allows them to perform the function that the Constitution of the United States uniquely outlines for them. The only profession called out in our constitution is journalism and its role to keep politicians in check. So we’re at an interesting crossroads now.

Eric Schurenberg (43:14.353)
Yes, yes. On the question of keeping politicians in check, one of the rating axes for the media bias rating that you produce is about reliability. And I’d like to come back to that because it does put you in the highly controversial position of determining who has a better grasp on truth and factuality among news organizations. That’s… difficult and we’ve seen how divisive that can be about who really gets to decide what’s true. How do you guys decide?

Vanessa Otero (43:56.692)
That’s a great question. And just because it’s difficult to do doesn’t mean we shouldn’t do it. But the difficulty is alighted by the fact that leaders of the major platforms, social media platforms, have said many times that we don’t want to be the arbiters of truth. But in our current moment where we have so much information out there and some of it, we know that some things are truer than others, right? Like we all get that basic concept. But then when you are looking at individual pieces of content to try to determine that, folks quickly get lost like, well, how do you know? How do you know what’s true? That’s a question we get all the time. How do I know if this is true? People can submit articles on our website and they’ll submit them from reputable publications. They’re like, I just don’t know. I don’t know what to believe anymore. And so many people have become apathetic because of that. So what we do, is we know that we can determine a likelihood of veracity for any implicit or underlying claims of pretty much any piece of content we come across. How are we able to do that? Well, all humans have the capacity to determine the likelihood of veracity of something they come across, right? We don’t often know the ultimate truth of really anything, right? And we just operate like we do. I mean, going back to, you know, early philosophy,

Philosophers have wrestled with how do we know what’s true, like even this world that we live in, how do we know for sure that we’re not living in a simulation of that, you know, and plugged into computers like in the Matrix. And the short answer is we don’t, we don’t know that for sure, but we have to operate as if, you know, this world is real and true because there are other consequences if we don’t.
So everything from that level of certainty that like this world is real on down is gradients. And we’re always constantly as humans making determinations of like what we believe is more credible than others, other things based on a bunch of different reasons. Now you can develop and hone those reasons in those faculties with practice and training. And that’s fundamentally what we do.

Vanessa Otero (46:20.404)
When we’re doing our evaluations. So there are basically four main ways people evaluate news content for how true something is. And some are better indicators, some are better heuristics than others. So if you think about the acronym RELI, R-E-L-I, there’s reputation, evidence, likelihood, and incentives, okay? People tend to over rely on reputation. Like I’m familiar with this, I know it.

Um, you know, it’s, it’s true. Uh, we also tend to over-rely on the last one. Incentives like, well, you know, the, um, the government wants to control me. So, you know, maybe they’re lying or this person is making money. So maybe, maybe they’re lying, you know, incentives are in reputation are kind of tough, uh, but they’re easy, they’re easy shortcuts. Um, evidence and likelihood are really the two main ones. And those are things that you can dig into. Um, by using the information that’s available on the internet. We have so much information available to us, but it takes looking it up to determine how much evidence there is for a claim or how likely a particular claim is. So what our analysts do is basically a combination of fact checking and other evaluations of the likelihood of veracity of content. So we rate things for reliability, even if they’re not. Quote-unquote, fact-checkable. Fact-checkable universe is just a subset of all content.

Eric Schurenberg (47:52.185)
Okay, great, thanks. I’ll remember the RELY acronym as a shortcut. I want to ask you just to conclude here. The company has been around since 2018. How are you going to measure success? How will you know that you’ve made a difference, that you have accomplished the benefit that is part of your charter, what are the measurements?

Vanessa Otero (48:24.34)
Well, as a company, we’re a public benefit corporation, which means we are for profit but with a stated public mission, which is to rate all the news to positively transform society. And people look at that like, what do you mean, how could you positively transform society by rating all the news? You know, we view news ratings, like just having markers of reliability and bias as a basic is a basic foundational tool to help news consumers, educators, business owners, social media platforms, all stakeholders in good news media be able to sort through our vast information ecosystem, which is getting bigger and bigger by the day. So as far as success for our company, look, we want to be a world renowned, the independent third party news, a rater of news. That’s what we want to be seen as. But what that looks like for the world if we have accomplished our mission is like a slow transformation that you don’t even notice at first. It’s younger people like growing up automatically knowing how to tell for themselves how reliable or how biased something is.

You know, they have TikTok today, they’ll have something else in five years from now, they can come across any piece of content, and they’ll know that there’s a difference between the reliability and bias. The demand for that outrage generating content will go down. And you’ll start to see that a trend, the big one is the trend in trust in institutions, including journalism will go back up.

We’ve been on this terrible downward every single pole. Down, down, down for trust. When we start to see that going up, that’s when you will be able to tell that we have made a difference.

Eric Schurenberg (50:23.161)
That is a great place to leave it. We can all devoutly hope for an upswing in trust in knowledge-based institutions, especially in the profession that has been so important, the fourth estate, and has been so good to me and that we all wanna promote here. So thank you. Vanessa, Lou, good luck with that, Fontes, and thank you for being on In Reality.

Lou Paskalis (50:49.07)
Thanks, Eric.

Vanessa Otero (50:50.132)
Thank you, Eric.

---

Created & produced by: Podcast Partners / Published: Oct 10 2023

---

Share this episode: