Trends In Newsrooms: The rise of the robots

Welcome to your automated news future, where low-value or formulaic content is produced by algorithms, not journalists. Jake Evans reports on the growing trend to automate report writing.

by WAN-IFRA Staff | July 6, 2015

The Associated Press (AP) has just hired an Automation Editor, a first for the industry. So, if nothing else, at least one journalist will get a job out of this. At the time of writing, the most recent news from AP on the automation front was that they would also be using automated reporting to cover thousands of college sports games that they didn’t previously report on.

“If people think this isn’t going to happen quickly – it is,” says Vice President and Managing Editor of the Associated Press (AP) Lou Ferrara who sits at the frontline of automated reporting.

If you don’t believe that journalism and society as a whole are about to face serious change at the hands of automation software, then consider these examples from The New York Times’ Human or Computer quiz:

“Kitty couldn’t fall asleep for a long time. Her nerves were strained as two tight strings, and even a glass of hot wine, that Vronsky made her drink, did not help her. Lying in bed she kept going over and over that monstrous scene at the meadow.”

–From the Russian novel “True Love,” written by a computer in St. Petersburg in 72 hours.

“A shallow magnitude 4.7 earthquake was reported Monday morning five miles from Westwood, California, according to the U.S. Geological Survey. The temblor occurred at 6:25 a.m. Pacific time at a depth of 5.0 miles.”

–An excerpt of an initial report by “Quakebot” about a March 2014 earthquake. The software took earthquake data, wrote a basic story and then woke up the journalist to fill the story out.

“When I in dreams behold thy fairest shade. Whose shade in dreams doth wake the sleeping morn. The daytime shadow of my love betray’d. Lends hideous night to dreaming’s faded form.”

–An excerpt of a poem written by machine-learning app Swiftkey after being fed a dataset of Shakespeare’s words.

Now that you’ve climbed back into your seat from the floor (seriously, that poem was written by a computer?), let’s examine where automation technology stands, and how it is poised to affect every newsroom in the business.

A software update: Automated reporting so far

The prospect of automated news became more real when AP announced last July that it would be fully automating its corporate earnings reports. This meant that from the sourcing of data to the publishing of a complete story, no human would be involved in the reporting process. A program written by the automation software company Automated Insights would pull data from Zacks Investment Research, sift it for “newsworthy” information, and then report this information in a way that is indistinguishable from human writing.

The stories that this automated reporting took over were simple, but the impact on the AP newsrooms was huge. The AP’s Ferrara says that not only did the software from Automated Insights put out 10 times the number of earnings reports for AP than before, but that it also freed up about 20 percent of his staff’s writing time.

“My staff, in general, is feeling like: ‘Wow, we have time we didn’t have before – it’s allowed us to do things we weren’t able to before,’” says Ferrara.

This development was hailed as the good news story that newsrooms had been desperately waiting for. Automated reporting was “the best thing to happen to journalists in a long time”, wrote New York Magazine’s Kevin Roose.

Bring on the goddamn robots,” called TechCrunch’s Alex Wilhelm.

The reaction to automation developments has been “giddy,” and any concern of future job losses has typically been waved away with reassuring statements like: “Nobody has lost a single job because of us,” in the words of Kristian Hammond, founder of automation software company Narrative Science – the main competitor of Automated Insights.

But despite the generally enthusiastic reception from journalists, critics such as Martin Ford, a leading commentator on job automation, MIT economists Erik Brynjolfsson and Andrew McAfee, and even Bill Gates, are worried.

Unemployment by tech: It’s different this time

“If you talk to the companies of these technologies – the same is true of companies who make robots – almost without exception they will claim that their technology is not going to displace workers. Nobody wants to be accused of that, so they always put a positive spin on it,” says Silicon Valley entrepreneur Martin Ford, a leading commentator on the potential impact job automation will bring to the economy. His latest book Rise of the Robots details how Artificial Intelligence is making “good jobs” obsolete – including journalism.

“Usually what they’ll say is ‘This just does the boring stuff, and it will free up people to do more interesting things,’ but realistically you know that that’s just spin. People will lose their jobs and they have lost their jobs already,” says Ford.

However, Automated Insight’s Vice President of Product Engineering Joe Procopio makes a valid point that as it stands, automated reporting is best-suited to number crunching and doesn’t have the capacity for the “how” and “why” of a story.

“We’ve done over a billion stories and we haven’t lost any journalists yet, and me having an affinity for that world, I can understand how the evolution of technology will impact the journalism career in the future, but I see way too many examples of journalists embracing where we are, and in some senses trying to be ahead of where we are,” says Procopio.

The threat to jobs is not something the AP’s Lou Ferrara agrees with either, since he believes that any work taken from journalists by automation will only give them new opportunities for deeper reporting: “I’ve talked to enough editors and producers and executive producers to know, that if they got rid of certain jobs, they would redeploy those positions into more important positions for the operation,” says Ferrara.

Procopio also argues that the work automation takes from journalists is “not so much a job to be replaced as it is a bunch of tasks that really don’t belong in the job description to begin with.”

“Maybe it will, in some markets, eliminate some jobs that are no longer needed,” says Ferrara, “The way printing presses changed, and pagination changed, and digital photography changed things over time in the news business; those jobs did, in fact, go away. But I would argue that over time the journalism industry creates other jobs of things we need that are frankly of higher value and more important, and I see that here too,” he says.

This Keynesian long-arc theory – which suggests technological unemployment is just a temporary phase of maladjustment – sounds rational, but unfortunately Ford believes that this time it doesn’t apply. “This time is definitely different, I believe that we’re looking at something really dramatically different than what we were looking at in the past [industrial and technological revolutions] – but it will take some time to really see that disruption,” says Ford.

Job displacement by this automation explosion is not news – as The Economist, MIT and Oxford have all noted – and it would be naive to think journalists are immune to automation.

However, there will be no “slaughter of all the journalists,” Ford says. All the classic woes of the journalism profession have had a much more dramatic impact than automating stories has, or probably will have, according to Ford. The problem is more subtle:

“I don’t think we’re going to see huge numbers of journalists lose their jobs because of this, but I think that what we’re more likely to see is fewer journalists may be hired in the future, and I think that’s especially true at the entry level.”

SportSport“If you want to have a career in journalism and you’re just graduating from school, the place you might start, your first assignment may be one of those routine formulaic things, like sport stories and corporate earnings reports and maybe even obituaries or something like that, you know, those kinds of things where there’s not a lot of creativity to it … that’s been the way that journalists have learned the ropes from the beginning,” says Ford.

Procopio disagrees with this claim: “I don’t see how the regurgitation of fact is something that a fledgling journalist necessarily does to learn the ropes, so much as it’s the garbage work that the experienced journalists don’t want to do.”

As far back as 2010, automation software was taking jobs from young journalists, like when Narrative Science software took over the recaps of baseball and softball games for The Big Ten Network – a job previously done by university sports offices.

“No newsroom is saying, you know, I don’t need as many reporters as I have,” Ferrara says. But while a company with the scale of AP probably won’t be using automated software to save a few newsroom jobs: “It’s like any technology, some people are going to use it for evil purposes, right: ‘Okay, I’m going to get rid of 100 jobs by doing this.’ But my goal remains to free up time and to free up resources to do the stuff that matters most.”

Getting personal: writing stories for the individual

Users of automated reporting software must tread ethical water when it comes to personalisation of content, and how far they might take the trend. And personalisation is something that Evgeny Morozov, an authority on the political and social implications of technology, indicated in Slate the software is well positioned to do:

“Imagine that my online history suggests that I hold an advanced degree and that I spend a lot of time on the websites of The Economist or the New York Review of Books; as a result, I get to see a more sophisticated, challenging, and informative version of the same story than my USA Today-reading neighbor. If one can infer that I’m also interested in international news and global justice, a computer-generated news article about Angelina Jolie might end by mentioning her new film about the war in Bosnia. My celebrity obsessed neighbor, on the other hand, would see the same story end with some useless gossipy tidbit about Brad Pitt.”

It is also one of Ferrara’s concerns as automated reporting develops. “The thing that I worry about the most as it relates to journalism is if you give people only what they want to hear, that’s not a good thing for anybody,” he says.

But Procopio doesn’t see personalisation being used this way. “I don’t think the draw of it is the personalisation or the contextualisation of fact for a single reader,” he says.

“Personalisation to me makes much more sense when you’re talking about, for example, if you were to do a story about stocks, their performance in the market today, we would personalise that based on the stocks in your portfolio, so you have the most relevant information to you at your fingertips at any given time.”

Procopio points out that personalisation already exists on an organisational level, as businesses cater to their demographics, and that automation is unlikely to take that to the nth degree.

“I don’t think rearranging the facts in those stories really buys an organisation anything,” he says.

A light at the end of the churnalism tunnel?

Before you take an axe to your laptop, consider that while automation – especially in its current state – will not completely free journalists from the constant 24- hour news cycle, Ferrara believes automated reporting “actually gets us something back.”

“In automating [news] we’re getting out fast and quick with just the facts, the basics, and I think for any news outlet to survive we have to produce content that nobody else has, and in order to do that you need time, and you only get that time by getting rid of the stuff that isn’t high value anymore. So I look at automation as much more a methodology by which to give people back time to do the stories of higher impact and higher value.”

“I’ve always looked at things that way – I oversee sports, entertainment and business news, and when I got into sports and entertainment, it was not lost on me that sports and entertainment are key news verticals that drive a lot of revenue for companies, and they help feed the larger journalism mission for the AP, there’s no question about that, and I see the same thing here. If this is able to solve problems and make customers happy and feed the larger mission, then that’s success.”

“I think the opportunity is there,” says Procopio, “To be able to update stories [automatically] as new data comes in, and certainly be available at any time.”

“Everyone talks about Quakebot, that’s an excellent example: not only did Quakebot sense that LA earthquake last year but it woke the journalist up, sent a message to his iPhone, got the journalist out of bed so that journalist could write the story to it. In that sense, it could allow the journalist to lead a normal life, so to speak, being made aware immediately when the data changes enough for it then to be necessary to do a rewrite, or do a new story.”

Automation pros and consAutomation pros and consYour automated future

The first AP Automation Editor started in March. “That person is actually designed to go through the AP and figure out what else we can automate. Because you can imagine after 168 years I’ve got a lot of processes in this place,” Ferrara says. “You’re going to see it more in the core operations of what all newsrooms do.”

However, Martin Ford is certain that for the rest of the industry, as automated reporting develops, it will only continue to stifle new jobs. “Whenever I have this discussion about robots and automation, people tend to have this idea that these technologies are going to arrive in huge numbers and people are going to lose their jobs right away. That’s not necessarily going to happen, but if over time we hire fewer people, if fewer jobs are created then in the long run that’s tremendously disruptive.”

It’s also a trend that won’t stay within the bounds of professional journalism. “Ultimately,” says Procopio, “I think it’s going to be very broad and that’s one of the things that we’ve been chasing, it sort of defines what we do. Eventually anyone will be able to create their own automated report, based on data they have.”

“Finding the facts is easy,” says Procopio, “It’s the prioritisation and the contextualisation of those facts that are difficult. That’s where I think [automation software] is going to evolve, especially in the next two or three years, and that’s where I see journalists evolving with it – having all that work done for you, and then as a journalist following up with the ‘how’ and the ‘why.’ ”

“You can’t dismiss the fact that it’s going to be more dramatic as the technology progresses,” says Ford. “Moreover, it’s not just automated content that’s evolving the technology – analytics, drones, sensors, crowdsourcing – they’re dealing with technological advancements on a day-to-day basis and automated content is just one of them,” Procopio says.

Whether you’re ready or not, automated reporting is set to make a big appearance in the journalism profession, and as it develops, to take larger and larger slices of the journalism employment and production pie.

Trends in Newsrooms 2015 

This case study appears in the World Editors Forum’s Trends in Newsrooms 2015, which is free for members to download. Read about the two of nine trends identified – Gaming the News – in part one of our #TrendsinNewsrooms 2015 blog series – and Source Protection Erosion in the Digital Age, the second trend featured.

Trends in NewsroomsTrends in Newsrooms

Trends in Newsrooms 2015 

This case study appears in the World Editors Forum’s Trends in Newsrooms 2015, which is free for members to download.

Share via
Copy link