Editor’s note: On May 30, Joe Procopio, who is vice president of product at Automated Insights and a WRAL TechWire Insider columnist,spoke on the future of Digital Journalism at Columbia University at a conference at the Tow Center. In a previous column, he discussed why automated content should be considered a tool and not a threat. Automated Insights focuses on auotmated content.

DURHAM, N.C. - I was prepared for just about anything when I walked up onto the stage at Pulitzer Hall at the Tow Center for Digital Journalism at Columbia University last Friday. I was speaking on the Robot Reporter’s impact on the future of journalism — the last session of the day at their first ever Quantifying Journalism conference.

In the four years that I’ve been designing and developing automated content at Automated Insights, the reaction of traditional journalists to our technology has been all over the map.

At one point it was straight-up denial – there was no way machines could create content, and even if they could, it would be painfully obvious that the content was created by a machine. See: Mad Libs.

Then there was fear – machines might someday be able to create viable content, which would put suddenly expensive human journalists everywhere out of a job, and maybe even start killing them. See: Skynet.

Fear drifted into a kind of bemused resignation – machines would indeed be able to create some kind of passable content, thus reducing a centuries-old dignified art to a constant stream of listicles and slideshows with soundbites. Since this was exactly where those same defeated journalists saw the industry heading anyway, it wasn’t too much of a stretch or a shock.

I had spent the entire trip from Durham to New York anticipating all kinds of questions – from the curious to the skeptical to the cynical. What I didn’t anticipate was a warm welcome, but that’s exactly what I got.

In fact, I can pinpoint the exact moment where I finally felt like the chasm between traditional journalism and automated journalism was bridged, for me anyway. It came at about 11 a.m., when Tow Fellow Alexander Howard, himself a columnist at TechRepublic, put up a picture of a robot typing away at a computer.

“The robot writers are already here,” he said to the gathered journalists and media executives. “And you should be prepared to work with them.”

No one ran from the room screaming. No one gasped. Many heads nodded.

Oh wait, I thought, not only do they already get it, they’re kind of ready for it.

More from Joe Procopio at WRAL TechWire:

  • Watching HBO’s Silicon valley – Yes, it’s good.
  • Rebooting the smartwatch and other wearables
  • Is ageism a problem in Triangle startups?
  • What’s up with Popcorn Time? What does it mean?
     

The conference was very well attended, with about 500 attendees and over 20 speakers representing The New York Times, Upworthy, The Pew Research Center, ProPublica, and various top journalism schools and huge media outlets from across the country.

The day kicked off with a panel on web analytics, as one might expect, what with advertising still easily bringing in the lion’s share of media revenue. But even here, there was an extended discussion of the validity of analytics in measuring the impact that a chunk of content has on the reader.

A lot of the time, I heard, no one really has a handle on that value. In fact, in some cases, they still rely on an editor to get an experienced “feel” for how an article performed.

Even with all those stats, figuring out the part that pays the bills is still very much gut. So right away I had a little cause for concern.

But then the conference quickly took a left turn into some cutting edge topics. Howard, the fellow who offered the warning to the luddites, went on next to discuss how data journalism was the latest step in its evolution and how journalists were evolving with it.

I was also pretty impressed by some of the advancements I saw throughout the day, especially around areas like algorithmic accountability, user-generated content, sensor and satellite journalism, and even something I’m just going to call maker journalism.

By the time I got up to speak, it was like the startup world had exploded all over the newsroom, and it was accompanied by a similar vibe, one of optimism. And excitement.

My remarks revolved around four items:

1) Who We Are

Once people realize there is such a thing as automated content, the next question is usually: Where can I get it? It’s a new science, and since Automated Insights will produce over 1 billion (with a B) articles in 2014, I feel confident claiming that we’re the leader in this new space.

2) What is Automated Content?

Yeah, so while a lot of people know what automated content is, still more think it’s something that it is not. It’s not templates. We’ve already gone beyond filling in canned sentences with the proper numbers. It’s algorithmic writing, and it’s about volume, personalization, localization, and speed.

3) For Most Journalists, Automated Content is a Tool, Not a Replacement

There are still quite a few things human journalists do today that automated content can’t. Like any advancing technology, automated content will eventually catch up. But as the technology evolves, good journalists will evolve with it.

In four years and after 300 million articles, we haven’t put a single journalist out of work. In fact, we’ve hired two.

This point, coming after a day of demonstrated advancements in the field, was not hard to get across. For a deeper discussion, see Why (Good) Journalists Have Nothing to Fear from Automated Content.

4) The Last Few Years Have Been About Perfecting The Robot Writer. The Next Few Will Be About the Robot Reporter.

The robot writer is fully baked. It’s a technology that we’ve made robust and versatile, figuring out how to incorporate tone, style, topic, prioritization of facts, flow, and lexicon. What we’ll be doing now is figuring out how to get more and disparate data into the engine. For more on this, see As Robot Writing Matures, Robot Reporting Is the Next Big Thing In Digital Journalism.

The questions, both during the session and for an hour-plus afterward individually, were also all over the map, but they were all about how the technology would expand and how it could be used. Here are three of my favorites:

  • What kind of people are automated content creators?

We figured out that people who are “just” journalists or “just” technologists aren’t going to be very good at training robots to write. This means there are new roles on the horizon, with a mix of skills that includes writing and editing, but also statistics, programming concepts, data science, artificial intelligence, and soft skills like pattern recognition and multi-dimensional thinking.

  • How do you promote accuracy with automated content?

Accuracy is huge for us. One of the sayings I use all the time is “We don’t always have to be great, but we can’t be wrong.”

On one hand, accuracy is a little easier with automated content because the data doesn’t lie. But that makes for a really dry “regurgitation” approach to the story. One of the things that makes us unique is that we apply aggregate, historical, real-world, or any other context we can determine.

This is risky, because we always have to know where the lines are between good and bad, positive and negative, certain and uncertain. Outlier identification is also a big issue. Most often, when we see a mistake, it has to do with either bad data or unexpected outliers.

  • How do drones fit into automated content?

I honestly don’t know how that would work today. Automated content has a human element to it, for sure, but there is so much human analysis required to interpret what a drone sees, at this point it isn’t structured enough to make broad use of the technology.

At the same time, sensors are really where it’s at for us. We’ve done traffic reports based on sensors, and as we start seeing more and more geological sensors, chemical sensors, that kind of thing, we’ll be able to make all kinds of new determinations.

So there are still plenty of technological areas in which automated content itself can evolve. As it does, there will be new opportunities for the cutting edge journalist to expand into new areas, cut time to press, and break new stories that can be told in new ways.

What’s inspiring is that the journalists are ready to make the most of the technology to do just that.

About the author: Joe Procopio is a serial entrepreneur, writer, and speaker. He is VP of Product at Automated Insights and the founder of startup network and news resource ExitEvent and new venture Teaching Startup. Follow him at @jproco or read him at Joe Procopio.com