Inside Steven Brill and Gordon Crovitz’s War on Fake News
The co-founders of NewsGuard, coming soon to a social platform near you, tell us why they will succeed where Silicon Valley has failed.
The role played by social media giants in facilitating the spread of deliberately false or misleading news reports on their platforms has thus far elicited several heartfelt apologies, but few inspired answers.
“It’s clear now that we didn’t do enough to prevent these tools from being used for harm,” wrote Facebook CEO Mark Zuckerberg in prepared remarks to Congress, released Monday. “We didn’t take a broad enough view of our responsibility, and that was a big mistake. It was my mistake, and I’m sorry. I started Facebook, I run it, and I’m responsible for what happens here.”
Suffice it to say that not everyone is convinced Facebook’s proposed solution—which, as of now, essentially boils down to serving users less content from third-party publishers and more content from their friends and family—will adequately tackle the issue.
A pair of veteran media execs believe they can help. The brainchild of Steven Brill and Gordon Crovitz, a new startup called NewsGuard aims to help educate consumers on the nature of the news sources they encounter online—specifically, applying labels to help distinguish legitimate news outlets from those that willfully aim to deceive people with misinformation.
Brill, who founded The American Lawyer and CourtTV, and Crovitz, a former publisher of The Wall Street Journal, say they aren’t out to become the arbiters of which outlets exhibit a conservative bias and which ones tend to adhere to the left. Rather, their targets are the blatantly deceptive—sites like the “Denver Guardian,” whose false stories, including one claiming an F.B.I. agent related to Hillary Clinton’s email disclosures had fatally shot himself, rode Facebook’s algorithm to appear in the news feeds of millions of users just days ahead of the 2016 presidential election.
Free to access and funded by licensing fees from social media platforms and advertising agencies seeking brand-safe environs for their clients, NewsGuard will provide “nutrition labels” to be placed alongside media outlets appearing in news feeds or search results, providing context around an outlet’s track record, ownership, or funding.
It’s not the first time a service has attempted to apply a human element to online fact-checking (look no further than Snopes or PolitiFact), but Brill and Crovitz say NewsGuard is different, not only because it will enlist a team of trained journalists to research some 7,500 news sites, but also because it will examine each outlet for its overall merit, as opposed to taking an article-by-article approach.
Armed with $6 million in seed funding (primarily from Publicis Groupe, one of the world’s largest advertising agency holding companies) and projected to launch before the 2018 midterm elections, NewsGuard’s biggest challenge will be achieving buy-in from Silicon Valley and the millions of individuals who rely on its platforms to get their news.
Folio: sat down with Brill and Crovitz to learn more about NewsGuard and why they believe it can succeed where the platforms have failed.
Folio: What about NewsGuard and its team of journalists will enable it to tackle the Fake News problem more effectively than the platforms’ own algorithms?
Steven Brill: I think you answered your own question. No algorithm can tell the difference between the Denver Guardian and the Denver Post if the Denver Guardian is trying to pose as a newspaper. Algorithms can do a good job identifying hate speech, because you can program in a bunch of words to look for. But they’ve found that it’s impossible to deal with fake news using artificial intelligence. If they could do it, they would have.
Folio: What types of traits or credentials are you looking for when you’re putting together your team of analysts?
Brill: It’s a mix of people with a lot of journalistic experience, such as [former Chicago Tribune managing editor] Jim Warren and [former Reuters news editor] Eric Effron, as well as some of the others who we are about to hire. The analysts drafting the actual nutrition labels will typically be journalists with two to four years of solid writing and reporting experience, preferably who also wrote for their college paper, as well as fact-checkers who unfortunately have been laid-off by some of the bigger magazines. It’s going to be people who have demonstrated reporting and writing skills.
Folio: When NewsGuard analyzes an outlet, how will that editorial process play out?
Brill: Two separate journalists working independently will rate a site either “green” or “red.” The first one will draft the nutrition label and the second will edit it, and add some cursory fact checking before our fact checkers handle the rest. In cases where the first two journalists don’t agree on the rating or if there’s any possible ambiguity on any of the 11 criteria, the supervising editors will make a decision, or Gordon and I will ultimately decide.
Folio: What might be an example of a disagreement that could arise?
Brill: Without getting into specifics, there might be a site where someone thinks their transgressions reach a level where they deserve a red, and the other person doesn’t see it that way.
Gordon Crovitz: Most of the 11 criteria are binary; clearly yes or clearly no.
Brill: You’ll also be able to click through to get the bios of the individual analysts who rated a given site. If you have some feedback or a complaint, that will be accessible. Additionally, if there is any box that isn’t checked “yes,” such as whether they’ve clearly labeled advertising as advertising, that won’t be there until we’ve called the site or emailed them to get their comment on it.
Crovitz: We’ll solicit feedback from readers, as well.
Folio: Are those roughly 7,500 sites you’ll be looking at confined to the websites of consumer-facing news outlets?
Brill: No, business and trade publications are included as well.
Folio: There has been some push back against fact-check sites like Snopes or PolitiFact over perceived political bias. Are you concerned about changing the mind of the reader who might be so jaded against the mainstream media that they don’t want the media to tell them what’s real and what’s fake?
Brill: I’m concerned about changing their minds; I’m just not 100-percent optimistic about doing it.
Crovitz: One thing that we think will be beneficial to everybody, including those who hold positions on the far right or far left, is that these nutrition labels we’re providing will give them more information about the sites that they’re reading. More information is valuable to everybody.
Brill: The nutrition labels answer some basic questions, such as: Who are these guys? Who is it that’s telling me this? If it’s about fracking, is it the American Petroleum Institute, or is it The Washington Post, or Folio:?
Folio: You had originally had a middle yellow category denoting sites that aren’t blatantly looking to deceive people, but have some influences that they aren’t necessarily disclosing. Why make the decision to do away with that?
Brill: About four weeks ago, we changed our minds and decided we were just going to do reds and greens. You have to be really bad to get a red. Otherwise, you’re going to get a green, but if you read the green nutrition labels you’ll see that not all greens are created equal.
Let’s say you have a site that does professional journalism, discloses its mission, discloses its financing, but it doesn’t label its advertising in a way that makes it clear to the reader that something is an ad. They still might get a green because that single thing is not enough of an offense to get a red, whereas something like RT, which does label advertising but doesn’t disclose its mission or its ownership and publishes all kinds of fake news—that’s going to get a red.
Folio: So would most of the sites that may have been a yellow under the three-category system now receive a green?
Brill: It’s probably about half and half. Some of them will be reds. The real test is whether they’re equipping the user to know where their information is coming from.
Folio: Looking at a platform like YouTube, where a new trending outlet could pop up seemingly overnight, how would your team handle quickly evaluating a new or unfamiliar source?
Crovitz: If there’s a new story trending online from a source we haven’t yet rated, we’ll have a “SWAT Team” of analysts ready to begin reporting on the nature of that source in real time. If you look back at what happened with sites like the Denver Guardian, we would have been able to report that those were fake news site within hours, if not minutes, whereas it took traditional journalism days or weeks to identify them, by which time a lot of their stories had already proliferated online.
Folio: So you want these outlets to take the ratings to heart and work to improve themselves?
Crovitz: Yes. The criteria include whether a site describes who is responsible for its content. Is there a way for readers to complain about mistakes? Is there a corrections policy? If a news site doesn’t get a ticked box on one of those, they’ll know they can improve their score by improving their accountability and transparency.
Brill: The platforms have said that they can’t share how they make judgments about what pops up first or what gets priority in a feed, because it would allow people to game the system. We want to be totally transparent because we want people to game our system.
Folio: It sounds like a lot of common sense; sites that are making a good-faith effort to produce legitimate journalism, to be transparent about who they are, probably don’t have a lot to worry about when it comes to their rating.
Brill: That’s absolutely true.
Folio: Platforms like Facebook and Twitter seem intent on giving off the impression that this is something they can handle internally. Do you expect them to get on board?
Brill: We wouldn’t have started the business if we didn’t. They want to see the product, and we’re going to be showing it to them fairly soon, but we are confident that one or more of them will launch by the end of the year. The proof that they’re likely to launch with us is that they still haven’t solved the problem. They’re paying for P.R. people and lobbyists to explain why it’s a really hard problem that they can’t solve, as opposed to using a tool like ours that can help solve much of the problem.
Folio: On the agency side, have there been any considerations you’ve had to make about who you’re willing to accept funding from?
Crovitz: We’ve posted the names of all of our investors on our website in the interest of transparency. You’ll notice that they’re not Silicon Valley companies or big media companies.
Folio: Is there anything else you’d want publishers to be aware of when it comes to NewsGuard?
Brill: We’re hoping that as we proceed, those publishers who do get greens will advertise the fact that they have a green rating. We think they should want to. That will help us establish our brand and it will help them maintain their brand.
Crovitz: We’ll have a way for publishers to use the NewsGuard logo and present our write-up about that brand if they choose to include that on their own website, which we hope they will.