To Free Themselves From Lies, Journalists are Using New Tools and Techniques to Check Facts

Posted

While the term “fake news” is thrown around with increased—and even disturbing—regularity, there remains the issue of whether reporters are doing enough to ensure that the information they write is factual and accurate. Today, reporters, as well as editors and the production staff, now must work faster than ever but getting the story out shouldn’t come at the expense of the facts.

Through the internet, stories can be posted almost as fast as news breaks, but the rush to be first can come with dire consequences.

“The fundamental flaw today with fact checking is that it is too often done after the fact,” said NewsGuard co-founder and co-CEO Gordon Crovitz. “Fact-checking often occurs today after the story has gone public, and this makes it very hard for the truth to catch up.”

More importantly, just because a story can now be more easily updated online to reflect changing “facts” doesn’t mean that fact-checking shouldn’t be as crucial as fact gathering.

Larry Parnell

“It is imperative that reporters ‘trust, but verify’ scoops, exclusives or other information provided to them so as not to be duped or used,” said Larry Parnell, associate professor of strategic public relations at George Washington University. “This may mean a follow-up call, double checking the alleged source or other traditional means of pinning down a story. Media relation professionals and reporters alike have to follow their ethical compass and not take shortcuts. Any such activity by the media will only magnify and support the claims that we are not objective professionals trying to tell a story fairly and accurately.”

The internet has also allowed for the reemergence of “fake news,” a type of yellow journalism that was once prevalent in the media. While the term has also been used to discredit legitimate news reporting in the modern day, there have been past efforts to stamp it out with notable success. This included the “Bureau of Accuracy and Fair Play,” which was started by Ralph Pulitzer and Isaac White at the New York World in 1913. It was focused on complaints and reducing the publishing of disinformation, but it also was meant to determine which reporters made repeated errors.

Fact-checking today remains a necessary part of the newspaper industry and for very much the same reasons as it did more than a century ago.

Empowering Journalists

Just as technology has made it easier for more people to regularly access the news, it has also made it easier for misinformation to spread. Thanks to our ever-connected world and notably the increased use of social media, which often allows information to spread unchecked, the dangers that misinformation presents is greater than ever.

Fortunately, there are now several resources journalists can rely on to determine fact from fiction, and this includes multiple online guides.

This includes the Public Data Lab’s “A Field Guide to ‘Fake News’ and Other Information Disorders” on hand, as it offers tips that can help understand the interplay between digital platforms, misleading information, propaganda and viral content practices. It is provides insight into the study of false viral news, political memes and trolling practices.

Moreover, there are an array of online tools that can be used to verify information online, and this includes mapping services such as Google Street View that can be used to pinpoint the exact location of a story to see if it matches with details from sources; weather data from sites including Weather Underground, which offers historical weather searches; and Snopes, a website that has been debunking urban legends and misinformation since 1994.

PolitiFact is leading the way on the political front by fact-checking statements, and this includes comments in speeches, news stories, press releases and even campaign brochures. While opinions aren’t checked, and there is room for “political rhetoric” and “hyperbole,” PolitiFact focuses on statements that are misleading and those that could be passed on and repeated by others. Efforts are made to be bipartisan, but PolitiFact also attempts to put more scrutiny on the party that holds power as well as those politicians that repeatedly make misleading statements.

“One error in reporting can hurt creditability with readers,” said Angie Holan, editor of PolitiFact. “The tools are important but the mentality to fact check something is critical. This includes the situational awareness that news that is gathered might not be factual.”

First Draft—a non-profit that supports journalists, academics and technologists—also offers several apps that can help journalists and fact checkers determine truth from fiction. These include a Verification Toolbox that lets users conduct source checks including social media presence, as well a guide for verifying photos and videos online.

Another tool is Hoaxy, which was developed to track the spread of information on social media, and this can allow reporters and fact checkers see how a story morphs as it spreads online.

Journalists should also be aware that even published stories can’t—or at least shouldn’t—always be trusted. FakerFact is a website that utilizes an AI to check the content of published stories, and it can determine whether the author was more focused on an agenda than the sharing of facts. However, FakerFact isn’t able to differentiate a quote from the body of the story and that could result in misreading and suggest bias that isn’t actually there. For those reasons, journalists should be prepared to do some good old fashioned digging and research.

“The primary tool that we can rely on is still something as simple as Google,” said Holan, who explained that searching published sources is now easier than ever, but she added that there are advanced searches as well. “People really need to be aware of how to use it for sophisticated searches. This includes advanced search for images as well as reverse image search.”

Checking to see if an image has been doctored, manipulated or otherwise presented in a misleading context has gotten easier. In addition to reverse image search on Google, there is also TinEye, a tool dedicated to determining if an image has been taken from somewhere else online, and that can allow fact checkers to determine if the image has been manipulated or modified. There are even TinEye browser extensions for popular search engines including Chrome, Firefox, Opera and Safari.

Checkmate

Non-profit Meedan has also released a set of tools that reporters can use during breaking news. Fittingly dubbed Check, an open-web based verification tool that relies on computer vision, natural language processing and machine learning/artificial intelligence (AI) to help journalists verify or debunk content shared via closed messaging apps.

Check allows users to upload social media posts, links, images and other content, which is tagged by subject matter. Other users can then add notes, other details and confirm or deny if the information posted in Check is accurate.

Megan Marrelli

“We believe that this will allow groups of reporters to collaborate to fight misinformation,” said Megan Marrelli, program manager at Meedan. “It brings people into an interface so that information can be verified more easily.”

Information gathering, including fact-checking, often begins with what you don’t know—but here is where past reporting can present problems for reporters and fact checkers alike. Just because something was printed in a news article or even published in a book doesn’t necessarily mean it was accurate.

“Very often there can be bias from past articles or published sources,” Holan said. “This is why a site like Wikipedia can be a great place to start, but you don’t want to finish there. Sites like this can be a good place to start your research if you don’t know the topic, but it is crucial when researching and fact-checking to go beyond it.”

However, these tools only take the role of fact-checking so far. Rolling up the sleeves and doing some sleuthing is still as important. That can include going to additional sources, finding published studies and checking past reporting. When talking to individuals it is also important to vet those sources and ask very specific questions. As Holan said, “Fact checkers shouldn’t be afraid to pick up the phone and ask a source if something isn’t clear.”

Gordon Crovitz

However, even “expert sources” shouldn’t always be taken at their word either. It is all too easy for sources to add personal bias whether it is intentional or not.

“Failure to disclose bias is a venial sin,” said NewsGuard’s Crovitz. “It is important for readers to know if a news organization is funding by a political organization.”

The State of Deepfakes

While misinformation has been spread for as long as the media has existed, a new and ominous threat has emerged—one that literally puts words into people’s mouths via online videos. Dubbed “deepfakes,” these videos first emerged in the late 1990s but now thanks to affordable video cameras and simple apps, anyone can make a fake video that seems all too real.

“History-wise, doctoring photos and videos is not new,” said Dr. Dustin York, director of undergraduate and graduate communication at Maryville University. “We’ve had doctored videos, even old school techniques that have been used recently. Deepfakes basically take that to the next step.”

Dr. Dustin York

As with some of the tools to fight misinformation, deepfakes are created by using machine learning to make a public figure say almost anything. It does currently need an existing video or facial scan. At present, it is something of concern for those running for office, celebrities or CEOs that are in the media a lot.

“Deepfakes give rise to concerns that increasingly advanced tools to manipulate video materials will also increase the capacity to manipulate people,” said Dr. Samuel Lengen, research associate with the Center for Data Ethics at the School of Data Science at the University of Virginia. “While we have had some time to get used to ‘Photoshopped’ images, deepfakes are a relatively recent phenomenon. One easy answer to the challenges it creates is that we need to raise media literacy and catch up with this new reality. However, the burden of responsibility cannot solely lie with the consumer.”

The other good news is that deepfakes still takes a bit of knowledge and expertise to create, and it also requires more powerful computers that what most users possess.

“However, it’s not the case where five people in the world only know how to do it,” York said. “It will become easier as the years go on. It still is advanced, but there are people who are capable of creating this content, and you can get it done if you find the right people.”

Fighting it will require efforts to monitor existing videos to find what could be a source for a manipulated deepfake video.

“You look for corroborating facts that will help to decipher between deepfakes,” said York. “Looking at calling out those very specific inconsistencies will help with spotting deepfakes and then outing deepfakes for public use.”

Fortunately, the tools to help reporters are on the way. This summer, the Washington Post launched “The Fact Checker’s Guide to Manipulated Video,” which includes three categories: Missing Context, which denotes if a video lacks or misstates the context of an event; Deceptive Editing, where by a video is called out after being edited or rearrange to change the context; and Malicious Transformation, where part of the video has been manipulated to transform the footage entirely.

Efforts to fight the deepfakes need to be as convincing as the videos. As York explained, “That’s how the public can build trust in those investigative journalists to weed through deepfakes when or if this becomes a tipping point of a real, major crisis.”

Preparing for the 2020 Election

The importance of fact-checking can’t be understated with the next presidential election now just a year away. Hard lessons could already be learned from Brazil’s election last year when Comprova, a coalition of 24 media outlets, worked to verify and debunk rumors that circulated online. This included tracking down where rumors and false stories began and working to debunk the misinformation. However, some candidates and political parties simply turned the tables and “denied” that what the fact checkers had found.

In Hong Kong, state media also used misinformation in the guise of news to discredit and undermine protesters, and this has included the manipulation of the context of images and videos.

This should serve as warning in the United States that fake news and misinformation isn’t used to influence voters.

“Here is where ‘misinformation’ is a loaded term,” said Marrelli. “Misinformation is actually a symptom of a much larger problem, and how to solve it is the big question. It can come from misleading articles and headlines, but it can come from those who are not journalists.”

Meedan is already working on its own initiative for the 2020 election, which includes efforts to find ways to better inform voters.

Lynn Walsh

“It is a complex problem that could have different solutions, but one way is to ensure that we accurately present where candidates stand on any given issue,” Marrelli said.

However, politicians often say things that aren’t true—sometimes it is an off the cuff remark, and sometimes it could be with nefarious motives in mind.

This is where FactCheck.org, the non-partisan, non-profit consumer advocate, and PolitiFact can work to respond to the comments made by politics and political candidates.

“Before we publish anything, we should double check what a candidate says,” said Holan. “The challenge now is again on social media. And it isn’t just from candidates but also third parties that ‘claim’ a candidate said something. That makes it even more challenging.”

Technology continues to evolve in ways that could be worrisome, in particular social media, which should be held to the same standards as the morning paper or five o’clock news.

“What happens is that people speculate on social media platforms,” said Lynn Walsh assistant director of the Trusting News Project. “You don’t want to say ‘I think it could be this,’ following a breaking news event. In other words, you can tweet out that there is smoke from a building, but don’t try to guess about what is causing the smoke. You don’t want to move forward with incorrect information.”

The Philippines could serve as a portent for what we could expect in 2020 as President Rodrigo Duterte was seen to use false stories, paid trolling and social media engagement—notably on Facebook—to win the election four years ago. This past summer, despite efforts to bring more transparency to digital campaigns, both Duterte and the opposition parties adopted new measures to throw off the fact checkers, while budgets for social media also increased.

“With every election there are so many things we’re not expecting, the dynamics change,” said Holan.

Chris Olson

There also remains the issue that with politics reporters can show partisan bias, and that is as damaging as misinformation.

Chris Olson, co-founder of privacy and security firm The Media Trust, said, “Fact-checking is extremely important to assure the messages that shape our world view and influence our decisions are based on facts. Democracy depends on an informed public. Once we are duped with false information, our democratic traditions will suffer.”

Comments

No comments on this item Please log in to comment by clicking here