• ADVERTISEMENT

    SXSW 2017: Embattled Journalism in Focus

    by Benjamin Bathke
    March 21, 2017

    The annual tech festival South By Southwest Interactive is widely regarded as an indicator of things to come across industries, and journalism is no exception. As a matter of fact, this year’s edition of SXSW was the first to have a dedicated “Journalism” track.

    “In 2017, journalism in the U.S. and around the world is under intense attack by various forces,” SXSW’s chief programming officer Hugh Forrest said. “Given the perilous nature of this industry, it makes perfect sense that we covered it extensively at SXSW 2017.”

    Like every year, dozens of media innovators and journalists descended upon Austin, turning the mecca for techies and creatives into a star-studded media gathering that included the New York Times’ Jim Rutenberg and Dean Baquet, former proprietor of Gawker Media Nick Denton, CNN’s Van Jones and Brian Stelter, Dan Rather, Recode’s Kara Swisher and many more.

    ADVERTISEMENT

    Of the 70 sessions in the journalism track, many were highly political, from the effect of filter bubbles and fake news on election outcomes, trolls’ role in the rise of anti-semitism online to President Trump’s antagonism towards the press. Other panels were about celebrity-driven headlines dominating the news cycle, the shifting role and responsibility of traditional media and redesigning online commenting.

    Here are three of the most timely journalism topics discussed at South By Southwest 2017.

    Automating Fact-Checking with Bots

    At a panel discussion about how technology can automate verification, four experts weighed in on the state of fact-checking, available tools, major limitations and what’s on the horizon: Alexios Mantzarlis of the the Poynter Institute’s International Fact-Checking Network, PolitiFact’s Bill Adair and Full Fact’s Mevan Babakar.

    ADVERTISEMENT

    The state of fact-checking

    Poynter’s Mantzarlis said although we are “nowhere near” automated live fact-checking, the current approach of tackling smaller problems, including trying to find claims to fact-check and figuring out how to fact-check simple claims, will “help make the process faster.”

    According to PolitiFact’s Adair, bots are not yet able to do complete fact checks by themselves but can help with individual tasks that “humans find repetitive, time-consuming and that might involve things like looking for factual claims to check.”

    Major hurdles and limitations for automated fact-checking

    • Nuance and access to data – Mantzarlis: “People make claims in all kinds of really complicated formulations that natural language processing is not currently capable of recognizing. The other challenge is access to data and comparing it across countries.”
    • Lack of data – Babakar: “More complex claims that require the combination of multiple data sets in a nuanced and logical way … is a very challenging task for a fact-checking system.”
    • Lack of resources, i.e. money and staff – Adair: “Fact-checking organizations typically are either legacy media organizations, radio networks or newspapers, many of which are having a difficult time financially,” he said, adding that independent groups like NGOs “may only have one or two staff members dedicated to fact-checking.”

    Available tools

    At Duke University’s Reporters’ Lab, Adair helped create iCheck, a tool that automates the time-consuming process of checking claims about a candidate’s voting record.

    Adair was also instrumental in developing ClaimBuster, a software developed at the University of Texas at Arlington that sifts through massive amounts of text to find factual claims to fact-check. “When we tested it, we found a really strong correlation between what ClaimBuster showed and what human fact-checkers actually checked from a debate,” Adair said.

    Flowchart detailing ClaimBuster process. Image courtesy of Chengkai Li, associate professor of computer science at the University of Texas at Arlington.

    What’s on the horizon?

    Unlike Poynter’s Mantzarlis, PolitiFact’s Adair is hopeful the automated and live verification of statements, such as instant “pop-up” fact-checking on live television, is only one to two years away.

    Adair admits, however, that we might see a pushback from TV manufacturers, especially when the technology is used to fact-check ads. “But we can create that same thing on an iPhone by holding it up to the television,” he said. “It detects a statement and pops up a fact check. I think that will be a game-changing development.”

    The Rise of Automated News

    Since news organizations like the Associated Press started using automated news software in 2014, robot journalism has been a growing presence in many newsrooms. The Washington Post, for instance, last year launched “Heliograf,” its own automated storytelling system, to cover the 2016 Olympics and the U.S. presidential elections.

    “There is a lot of interest from our editors and reporters in how Heliograf … can remove rote and repetitive work,” Jeremy Gilbert, the Post’s director of strategic initiatives, told me over email. Heliograf would allow them to focus on “more interesting and valuable reporting,” he said.

    Another strength of automated news, according to Gilbert, is its ability to personalize and customize news. To that end, he said, the Post will launch “new tools” this year to both help their reporters “identify potential stories and personalize stories” for the Post’s audience.

    Joe Procopio, chief innovation officer at Automated Insights, a company producing software that helps publications like the AP generate data-based content, said personalized journalism is especially beneficial for local and hyperlocal news sites as well as blogger whose constituents have a particular interest in local news, from sports teams and traffic figures to education and crime.

    Joe Procopio, chief innovation officer at Automated Insights. Image courtesy of Procopio.

    “Anywhere where there’s data that’s accessible, they can write stories about it,” Procopio told me in Austin. “In theory, just one person can run an entire newsroom in a small town with automated content.”

    Can automated news save local journalism?

    Although he doesn’t consider it a silver bullet, Procopio believes automated news can slow the demise of local news, both newspapers and online publications. “I believe it could allow the continuation of a smaller newsroom, for instance by providing more news to be able to fill a day,” he said. This, in turn, would make a publication “more full fledged” and help it attract more readers and local advertisers.

    Gilbert, on the other hand, is skeptical about automated news helping cash-strapped media outlets. “Automation is not a solution for all stories or a panacea for reduced newsroom resources,” he said, adding that it cannot replace human news judgement. “The machine is only as good as the journalists and technologists who power it. The world is not calling for a lot of simplistic, templated stories,” he said.

    To Gilbert, it isn’t a question of “humans versus automation,” but how “automation can help humans.” He said the Post will start using Heliograf for the automatic and continual updating of “living stories” so they are current “whenever someone arrives at the story.” Procopio calls this automatic actualization whenever new data is available “interactive journalism.”

    Where’s automated news headed?

    Gilbert, who participated in the same SXSW panel on automated news as Procopio, believes the process of creating automated content will get easier. “Non-technical journalists should be able to craft stories without needing to know how to program. Once that happens, it will be easier and less costly to automate. That’s critical to ensure widespread adoption of automation in newsrooms.”

    Procopio agreed, saying that thanks to automated news, journalists don’t have to be “data scientists” to produce data-driven content. He also pointed out that once widespread concerns, including robots putting journalists out of work, have given way to “much more acceptance,” especially over the last three years.

    “Five years from now, automated news will be a standard part of the reporting process,” Procopio predicted.

    Live Video: Pitfalls, Best Practices and the Future

    Two years ago, Periscope and Meerkat launched at SXSW. Since then, most big social sites have brought out their own solution for user-generated content livestreaming — Twitch IRL, Facebook Live and Instagram Stories have launched, Periscope is now integrated into Twitter and YouTube Live is continuing to evolve.

    In Austin, a panel of practitioners discussed the rapidly evolving live video space: Darian S. Harvin, editor and news curator at BuzzFeed, Deborah Acosta, interactive video journalist at the The New York Times, Mark Jones, head of digital content at the World Economic Forum, and Kay Meseberg, head of VR/360 at public Franco-German TV network ARTE. Here are the main takeaways.

    State of live-streaming

    • BuzzFeed’s Harvin: “We are in the very hard and fast, but still very experimental phase of live video. Publications are very much in a very aggressive stage right now seeing what works, noticing what their audience wants and pivoting from there.”
    • Times’ Acosta: “Broadcast news organizations have been doing live video for decades. What’s new about live video today, is that you can now interact directly with the audience while you’re live.”
    • WEF’s Jones: “The technology is very crude, and the editorial formats are elementary.”

    Potential pitfalls

    • Harvin: “Not get your reporters to dabble in this field, because the more you can assimilate this into your coverage … the better off you’re going to be in future because that’s where your audience is.”
    • Acosta: “Be careful about how you provide context when interviewing somebody with a particular opinion. You do not want to mislead your audience or come to conclusions early in the reporting process.”
    • Jones said it’s the “social video dilemma:” For reach, 60-second videos with big captions work best, he said. Facebook Live is the better if you want interaction, although it tends to have lower reach and possibly even a “detrimental effect on your overall Facebook rating.” He admitted he doesn’t know how to achieve both reach and interaction.

    Best practices

    • Harvin: “The more interactive you are with your audience, the better,” she said. “When you live-stream, you are able to break that fourth wall between the anchors and the audience.”
    • Acosta: “Answer your audience’s questions and respond to their comments in real time. And ask them questions as well.”
    • Jones: “The best combination is a personality with a strong social network on an interesting topic to your chosen audience and the talent’s willingness to share the live video with their network.”
    • Meseberg: “Elaborate 360-degree video live-streams with high production value only make sense for special occasions.” Most live events, he said, are better suited for “single-camera shoots that are quick and interactive.”

    Meseberg gave one example for each scenario, both of which were live-streamed on Arte’s YouTube channel: the opening concert of Hamburg’s new philharmonic orchestra earlier this year as well as two electronic music performances during SXSW Music.

    Ways to improve live videos

    • A tool that brings more context to your live video – Harvin: “Changing graphics, pinned captions or sending more updates while live-streaming.”
    • A way to curate comments afterwards – Acosta: “When you had a very popular livestream, you can get tens of thousands of comments. It’s hard to sort through all of that information. If there were a tool allowing journalists to quickly filter through all these comments and find little nuggets of wisdom, that would be really helpful, especially when you’re trying to put together a more polished piece from that live video content.”
    • Being able to do multiple things from your phone simultaneously – Jones: “Stream to main social networks, add in relevant contextual information including graphics and stills, and get people to share simultaneously.”
    • Live-streaming events in multiple languages and/or with subtitles – Meseberg said Arte is not yet actively working with such tools because the “quality is not sufficient today.” One tool that already offers live translation, however, is Skype Translator.

    Correction: The last paragraph of this piece has been corrected to better reflect that Arte is not currently working in the live translation space.

    Benjamin Bathke is an entrepreneurial freelance journalist covering media innovation, startups and intractable global issues for Germany’s international broadcaster Deutsche Welle, as well as several other international publications. In 2015-16, Ben was a Global Journalism Fellow at the Munk School of Global Affairs and a multimedia storyteller for Washington University in St. Louis. You can follow the 2017 Reynolds Fellow on Twitter and see more of his work here.

    Tagged: Automated journalism Benjamin Bathke fact-checking live video livestreaming south by southwest sxsw 2017

    Comments are closed.

  • ADVERTISEMENT
  • ADVERTISEMENT
  • Who We Are

    MediaShift is the premier destination for insight and analysis at the intersection of media and technology. The MediaShift network includes MediaShift, EducationShift, MetricShift and Idea Lab, as well as workshops and weekend hackathons, email newsletters, a weekly podcast and a series of DigitalEd online trainings.

    About MediaShift »
    Contact us »
    Sponsor MediaShift »
    MediaShift Newsletters »

    Follow us on Social Media

    @MediaShiftorg
    @Mediatwit
    @MediaShiftPod
    Facebook.com/MediaShift