The Robots Are Coming!

by Andres Oppenheimer

  When it comes to the way news is written, Gilbert confirmed that—just as journalists have been using spell checkers and grammar correction software for some time now—a number of new tools will come into play in the coming years to help us do our jobs. “More stories will be hybrids of humans and automation,” he said. “Software systems will be able to suggest different sources. They’ll be able to write and rewrite background and contextual information that reporters would otherwise be doing. They’ll be able to help identify tips for new stories. I also think that growing numbers of reporters will find themselves looking at their beats and finding ways they can expand the coverage they provide by using automation. So it’ll be the reporters who will be driving what stories we automate instead of editors or technologists, which is the way things are done right now at the Post. I already have a lot of people in the newsroom coming to me and asking, ‘Can’t you just automate this part of my job so I can do this other thing that I really want to do?’ ”

  Most data-centered sports and financial news will be automated, Gilbert continued. “If all you’re doing is looking at the data coming in on a corporate earnings report and writing about that, that’s not a particularly interesting story for anyone. It’s a story that a machine can do faster and better. On the other hand, if you’re seeing some unusual patterns in a corporate earnings report and you want to investigate what’s happening, that’s something a human can do much better than a machine.” And there are other kinds of stories, like obituaries, that are always going to need a human touch. A computer can look up a dead person’s résumé, draw up the facts of his or her death, and list the different positions that person held, but it won’t be interesting without a journalist to interview the relatives and tell stories that don’t appear in any official records, he explained.

  I asked Gilbert whether journalists at the Post’s newsroom see him as the enemy who sooner or later will automate all of their jobs and leave them unemployed. “I don’t think so,” he said. “I’ve been very happily surprised by the reaction to the automation tools, especially. There are lots of more people who are interested in how automation can help them than there are people actually working on automation projects right now. We’ve been lining up a whole host of potential users, and that’s a good thing. I mean, it suggests to me that the newsroom is much less afraid than they are eager to save themselves time so they can go and do other things. We ask a lot of our reporters and editors. We ask them to be active on social media, we ask them to help with the production of their stories, we ask them to be very active on television and to be quoted by other media organizations. If they can save a little bit of time, if they can identify stories that nobody else has, that’s definitely to their benefit.”

  THE WASHINGTON POST GOT A NEW LIFE THANKS TO TECHNOLOGY—OR THANKS TO TRUMP?

  Perhaps because it was bought by the megabillionaire founder of Amazon, The Washington Post has been one of the few newspapers in which automation has not resulted in massive layoffs. On the contrary, the paper—which like most papers in the country was increasingly losing circulation and advertising back in 2013, and had just laid off fifty-four administrative employees—was reborn that year after being bought by the Silicon Valley tech mogul. The Post doubled its staff of technologists, bet everything on its digital edition, and soon enough began to see a serious increase in readership. In 2017, it announced plans to hire as many as sixty new journalists, something unheard of in an industry that was in rapid decline. According to the Post’s executives, technology wasn’t destroying jobs, it was creating new ones.

  “We’re adding dozens of journalists,” said Fred Ryan, the Post’s publisher and CEO, in late 2016. Thanks mainly to technological advances, the paper has seen a 75 percent increase in new subscribers, and it has doubled digital subscription revenue over that same year, he said. It’s also getting ready to boost its mobile video output, send out stories via email, and promote investigative journalism. “Investigative reporting is central to our DNA,” Ryan said. “Readers expect it.”

  According to Politico.com, Bezos, the Post’s new owner, invested up to $50 million in the paper, mostly in technology improvements. Now there are roughly eighty technologists in the newsroom, along with over seven hundred journalists. “This is the face of a modern newsroom,” Politico wrote, “in which software development engineers, digital designers, production managers, mobile developers and video engineering produce content in real time.” And the investment is paying off.

  But to what extent was it technology—and not the barrage of news and scandals coming from Trump and his chaotic administration—that revived the Post in 2016 and 2017? Clearly, Bezos’s investment has allowed the Post to introduce new technologies and increase its readership by giving people news about local elections and high school football scores. But it’s also true that Trump’s unexpected rise in the polls during the campaign and his subsequent—and controversial—electoral victory have produced a hunger for news that has rarely been seen in recent U.S. history. And The Washington Post, like The New York Times, was well positioned to take advantage of that hunger for investigative journalism.

  THE DANGERS OF MICROTARGETING NEWS

  After interviewing some of the top technology gurus in the newspaper business, I have no doubt that the news will become increasingly personalized and microtargeted. Just as algorithms for Amazon and Google are already tracking our personal interests and bombarding us with ads tailored specifically to us, the same thing will happen with the news. Many people are afraid—and rightly so—that the widespread use of microtargeted news based on the economic status of the neighborhoods in which we live and our online reading habits will turn us into an increasingly fragmented society. We will be living—even more than we already are—in information bubbles. Theoretically, the news we get will depend on who we are and what we want. But that can lend itself to dangerous distortions.

  One of the media analysts I talked to who is most concerned about the dangers of personalized news is John Bracken, then the director of media innovation at the John S. and James L. Knight Foundation, which researches and finances new technologies for the newspaper industry. According to Bracken, “In the twentieth century, the notion of journalism and newspapers was the forum for an entire community. But what we’ve begun to see in the last twenty or twenty-five years is the disintegration of the mass media and a growing tendency toward microtargeted news.” What we’re seeing now is “the loss of a common public culture” and the formation of a “5,000-channel universe,” he added.

  The individualization of the news can leave us open to being politically manipulated, because the algorithms used by platforms like Google and Facebook are designed to satisfy the consumer, rather than fulfilling a public service, Bracken said. So what they do—as we saw with the 2016 U.S. presidential election—is to reinforce the political preferences of their audiences instead of giving them news from different angles so they can form their own opinions. In other words, Bracken said, the danger is that these technologies are encouraging extremist views.

  “If I’m Facebook, my mission isn’t to strengthen democracy. My mission is to keep you on my site and engaged,” Bracken told me. The Facebook algorithm knows our preferences by having studied our reading habits, and based on that knowledge it will try to please us by giving us information that keeps us happy and plugged into its platform. If it sees that you tended to read articles favorable to either Hillary Clinton or Donald Trump during the campaign, it’s going to give you more news that are in line with what you’ve read before. It’s not going to put any effort into offering you opposing views that will upset you and drive you off the platform, he explained.

  “I’ve seen a lot of attempts to build alternative networks, but people aren’t going to them,” he continued. “The reason Mark Zuckerberg is one of the richest men in the world, if not the richest man in the world, is because people like h
is platform and the way the algorithm works. It feeds people’s neurons.” Bracken concluded, “People, especially online, rarely say, ‘Okay, I need to act civically now and get really good, reliable news and information.’ When there’s so much cotton candy out there, people go after the cotton candy.”

  FACEBOOK AND GOOGLE: ON THE DEFENSIVE

  After Trump’s victory, and the revelations by the CIA and the FBI that Russia had tipped the scales in Trump’s favor by planting fake news on Facebook in key parts of the country, Zuckerberg’s first reaction was to reject the backlash and say his company didn’t generate news content but was simply a platform through which flowed all sorts of news. But when faced with a flood of criticism that Facebook’s algorithms were allowing—if not actively fueling—the spreading of fake news, Zuckerberg was forced to admit that something wasn’t working.

  In a public posting to his own page on November 19, 2016, he wrote: “A lot of you have asked what we’re doing about misinformation” and that “we take misinformation seriously….We’ve made significant progress, but there is more work to be done.” Zuckerberg promised to launch new ways of eliminating fake news, which included new algorithms for detecting made-up stories and tools so readers themselves can put up flags when they see false information being spread. Facebook also looked for help from external, nongovernmental groups when it comes to verifying stories and instituted new programs for preventing algorithms from promoting ads for fake news sites.

  But in 2018, Zuckerberg found himself in deeper trouble, after it was revealed that Facebook had failed to prevent Cambridge Analytica—a data-mining firm that had worked for the Trump campaign—from getting private information on 50 million Americans before the 2016 elections. Facebook’s stock plummeted, and Zuckerberg was summoned to testify before Congress. His renewed apologies did not convince many critics. Whether it was because of privacy violations, distribution of fake news, or growing complaints that they were intentionally creating “social media addicts,” Facebook and other social media firms were increasingly the object of a backlash against big tech, which is likely to increase if the trend toward personalized news continues.

  THE REINVENTION OF THE NEW YORK TIMES

  Not long after I spoke with Jeremy Gilbert, the director of strategic initiatives at The Washington Post, I interviewed Kinsey Wilson, who oversaw innovation and strategy at The New York Times. Wilson was promoted to executive vice president for product and technology in 2015. Under his tutelage, the paper began focusing its resources on the online edition and putting a great emphasis on the immediate publication of stories, promoting them on social media and increasing their visual content.

  In recent years, the Times had reduced its workforce because of the drop in advertising that affected all print media. In 2014, the paper had cut one hundred newsroom positions, and in 2017 it announced a new round of buyouts, mainly in editing and supervision jobs. “Our goal is to significantly shift the balance of editors to reporters at The Times, giving us more on-the-ground journalists developing original work than ever before,” the paper’s senior editors wrote in an internal memo. But at the same time, the online edition was growing by leaps and bounds, reaching 2.2 million digital subscribers and reviving the company’s hopes. The newspaper was being reinvented before our very eyes.

  Wilson told me that the Times, like The Washington Post, was experimenting with sending personalized news to readers according to their geographic location, and it was also successfully using artificial intelligence to manage comments from readers in its online edition. Due to the large number of insulting, racist, and poorly written comments, along with automated message chains sent by political parties or public relations agencies, monitoring readers’ comments had become a very labor-intensive task. Online “moderators” or human editors had to check the comments and then confirm whether they were authentic or sent by bots. This got to be so much work that many papers shut down their online Letters to the Editor sections in recent years. Instead, the Times made a deal with Google whereby the Silicon Valley–based giant started to handle the paper’s readers’ comments. The Times gave Google its entire archive of readers’ comments that had been fact-checked and edited by human journalists, and Google then used the data to create an algorithm that learned from the editors’ decisions and automated the process going forward.

  According to Wilson, the automated system has been working quite well. Thanks to the archive of the Times’ edited reader comments, the algorithm “did a very, very good job of predicting which comments were likely to be approved and which ones were likely to be rejected,” he said. As a result of that, the Times was able to increase its number of online stories with readers’ comments from 10 percent to about 80 percent. “In essence, it allowed us to substantially scale up the community engagement on our pages,” Wilson told me.

  THE MEDIA OF THE FUTURE: INTEGRATING TEXT, AUDIO, VIDEO, AND IMAGES

  When I asked Wilson how he imagines the future of journalism, he told me it will involve using interactive graphics, text, audio, and video to tell stories in an integrated way. Until recently, each of the media had its own way of conveying the news: print media did it through text, radio through audio, and television through video. But now there will be a new way of delivering the news, which will use all these communication forms and will drive us to use all our senses at the same time.

  We will be able to combine all those different elements into more powerful forms of storytelling, Wilson told me. “We’re moving from an era when television, newspapers, and radio were very distinct entities to one where you’re seeing a much more profound integration of all those different forms, resulting in the very best kind of storytelling that’s being done today. That, I think, is one of the trends we’ll certainly be seeing more of.” As an example, Wilson cited the Snapchat Discover application, where people can see videos combined with text, images, and even animations, in a very creative new way of presenting the news. Augmented reality will certainly be part of the media mix: The New York Times and The Washington Post have already used augmented reality to illustrate their coverage of the 2018 Olympic Winter Games in Pyeongchang, South Korea. Readers could type a code in the newspapers’ Olympic stories on their smartphones, and see close-up details of what they were reading, such as a nearly microscopic view of U.S. speed skater J. R. Celski’s gloves with plastic caps on his fingertips to protect his hands during turns.

  Wilson is also predicting that virtual assistants like Alexa or Google Home will become “one of the most significant consumer product developments since the introduction of the iPhone.” These voice-activated devices are already in our homes, and soon enough we’ll be using them to open doors, turn on lights, control our screens, and also deliver the news in a smart way, so we aren’t forced to read or listen to the same stories time and again.

  “You’re increasingly going to expect these devices to be smart enough to know what you’ve already read or seen or heard on the news,” he said. “At the end of the day, these devices are going to be smart enough to know what your cadence is, what your rhythm is, and they’ll be presenting the most relevant things they can to you.”

  RADIO WILL CONTINUE TO BE VERY IMPORTANT

  When I asked the New York Times’ innovation chief if radio has a future as an autonomous news delivery channel, I was surprised to hear him respond with a resounding yes. Wilson argues that radio has survived in a world dominated by television and has taken a long time to be influenced by the Internet because “it’s the only medium that’s at least partially passive. In other words, I can do other activities while I’m listening to radio. Television and print require 100 percent of my attention, or close to it, but with radio, I can easily move back and forth between full engagement and 50 percent engagement, depending on what other activities I’m doing. That’s why radio is so popular in the car. I can focus my attention on driving and be listening to the radio
at the same time.”

  But won’t all of that change in coming years, I asked, when self-driving cars become popular? “Yes, but not exactly,” Wilson replied. “Driverless cars will present a challenge for listeners, but I don’t think radio will be going away. It all depends on the quality of the production, the kind of experience you’re looking for. In many ways, nothing is more powerful than the spoken word in terms of getting inside people’s heads and invoking a certain sense of connection and emotional resonance. I think, in one form or another, radio will remain a very important, very durable thing.”

  Wilson added, “I do think, particularly with the introduction of smart speakers (such as Alexa or Google Home), we’ll see something similar to what happened in the music space over the last fifteen years: radio news reporting will gravitate quickly to these devices.” In other words, just as the Internet paved the way for the creation of Pandora, Spotify, and other platforms where we can listen to all sorts of music, more radio news broadcasts will be channeled through new platforms in our virtual assistants through the web. It won’t technically be radio, but in practice it will be the same thing: orally delivered news.