Bots and artificial intelligence arrive on the scene to help media engage with audiences


Watch video of the conversational journalism panel discussion.

When it comes to including robots in journalism, the greatest fear revolves around how they will replace humans in their reporting. However, if there is something shared by speakers in the panel “Conversational journalism: how bots and artificial intelligence can get us there,” it’s that this technology was created to help journalists in their work and that they are reliant on humans to operate.

The panel, which took place April 21 during the first day of the International Symposium on Online Journalism (ISOJ), was an opportunity for representatives from Quartz, The New York Times, The Washington Post and Condé Nast Austin to share their experiences creating bots for journalism and the use of artificial intelligence.

Speakers at ISOJ showed how they are experimenting with bots and artificial intelligence for journalism. (Mary Kang/Knight Center)

Sanette Tanaka, senior product designer at Vox and moderator of the panel, started the presentation by calling 2016 the year in which most media outlets said “let’s build a bot,” making 2017 the year with an innumerable list of tools made in the previous year.

For Tanaka, the most important part of the last year was precisely what bots have managed to achieve in newsrooms, leaving leaders of the technology to experiment with how the outlets can achieve better communication with audiences. However, as in most experiments, they need to prove their usefulness in practice, and that is where the challenge lies for 2017.

Tanaka shared her team’s experience creating an Amazon Echo bot and the question she asked about the value it would bring to their audience.

“Were we creating something that truly enriched our audiences’ lives? Were we creating Samantha or were we creating a somewhat self indulgent project that was cool and interesting to us  but perhaps somewhat forced and baffling on our user’s side?” Tanaka asked. “And the truth is, we probably created something in between and I think that’s where the challenges lies for us in 2017. How can we move from simply existing in this conversational space to truly creating value for our audience?”

John Keefe, bot developer and project manager at Quartz Bot Studio, agreed and said the main thing is to create a bot that can help journalism. According to Keefe, the Bot Studio at Quartz is experimenting with artificial intelligence as well as “developing tools and learnings for you [journalists]. We have a grant from the Knight Foundation to exactly that. This is not just for Quartz to learn how to do this, it’s for Quartz to learn how to do it and then share it with you,” he said.

According to Keefe, they are currently offering a bot through Slack for media outlets to use.

For media like The New York Times or The Washington Post, the use of bots to achieve better interaction with the audience has become a good way to experiment.

Andrew Phelps, product director at The Times, explained several of these projects including one that was put into practice to cover the Rio 2016 Olympic Games. The the paper’s sports deputy director, Sam Manchester, was responsible for reporting through SMS with readers who registered to received notifications.

“[They were] personal observations of the Games, not headlines, he should talk to people the way to the way he would talk to friend,” Phelps said.

However, it was not the only experiment. During the presidential campaign, political correspondent Nicholas Confessore also expressed his personal opinions about the election, which were connected to the newspaper’s political bot. People interacted with him through suggested questions.

But the most important thing for Phelps is the reaction of people who felt that it was not a robot- it had personality and managed to connect with the public.

“I’m not worried about this technology driving humanity out of journalism. I’m very excited about the promise of technology bringing more humanity to journalism and creating more one-to-one experiences with more people that otherwise might be impossible,” Phelps said.

The Post had similar experiences. According to Joey Marburger, director of product at the Post, they have tried to experiment with a lot of bots internally. One of these three bots has been released to the public in search of learning what people want. For this, they have created the three laws of journalism with bots which they use as a guide for what they should do as a newspaper.

Conversational journalism (2017)
L to R: Sanette Tanaka, Travis Swicegood, John Keefe, Andrew Phelps and Joey Marburger (Mary Kang/Knight Center)

But like Phelps, Marburger highlights how these bots are moderated by humans. For him, the most important thing is to move on to an experience that gives much more, and for that reason, they will continue to experiment.

He talked about how bots can fill spaces between platforms, like notifying you where you left off on a story. His team is looking into more applications for bots like this.

“I remember bringing this [bots] up in the newsroom, nobody really understood it, why would we do this? Especially when you do the first one and it gets like five people that use it. You’re like, ‘We gotta keep doing it. We gotta keep doing it. It turns out you learn a lot from experimenting,” Marburger said. “We’ll be iterating on bots for a long time to come.”

Travis Swicegood, director of engineering at Condé Nast ATX, focused on the need for media to understand that their content needs to be ‘rethought’ before turning it into a bot.

“The key thing here is to reframe how you think about the content you have,” said Swicegood, who added that the knowledge and content that media have cannot simply be passed to a bot. He explained that narrative is not the same as conversational and that you have to think differently if you want to make content more conversational.