By Steve Outing
Instead of writing a traditional text article, wouldn’t it be great to write in a way that you told your story by having a conversation with each person in your audience? (I won’t answer that just yet.)
Recently I ran across the intriguing mobile app and storytelling platform Massively. It can be defined as: online and mobile content + interactivity + mobile text messaging.
What?! Text messaging is an old technology. Could it be part of future forms of storytelling, or news? … Maybe.
I’ll explain Massively to you using the authoring platform. The web interface below is the published version of my first experiment using Massively. (It’s a web version of what you’d normally experience on a smartphone or other device.) Play along and you’ll learn more about the app and about its possibilities for journalism use.
(If the embedded player below isn’t working for you, try this direct link – http://goo.gl/R0h8nF )
It was very easy to produce the text-messaging “story” above; the Massively authoring tool was simple and quick to learn. To compose a story, you write along a decision tree. Based on users’ answers to questions you pose, they’ll see different components along the story arc.
My example above is pretty basic, but imagine a more thought-out version applied to telling a news story in this interactive way, where the user receives information via text chat with the (automated version of the) reporter.
An example might start out with a news text alert sent to subscribers:
"An explosion has leveled the American Bank building in downtown Seattle. Want to know more? ..."
Assuming that a user responds with a yes (or related word, which the Massively system recognizes through natural language processing), the answer might be:
"So far, we know about injuries, deaths, suspect under arrest. What can I tell you?"
Who did police arrest?
Since this is a piece of information currently in the story, Massively would interpret the user’s answer to be a request for information about the arrest. It would pull this answer out of the script that the reporter previously had entered into the authoring tool:
Police have arrested 35-year-old Joe Smith of Tacoma. He is believed to have been recorded by security cameras placing a bomb in the basement parking area. ...
Smith is being held in solitary confinement by the Seattle office of the FBI, who are involved in the investigation. ...
Do you want to know anything else?
And so continues the conversational news exchange. If a users asks for information that is not known yet (and thus not included in the script that makes up the text-messaging story presentation), the user might see:
No information is available about that yet. More? ...
Massively claims to be using “artificial intelligence” to interpret users’ text, so that it can deliver the piece of information the user is seeking. It’s not perfect yet, but the system does a decent job of understanding what the user is seeking without requiring that only specific words be used.
Now, I wouldn’t suggest this as a primary format for news presentation, but as a better alternative to simple news alerts sent to mobile devices, Massively’s model is a step in the direction of a more interactive future for news. It’s easy to imagine that as artificial intelligence advances and is applied in this manner, such a system might only need to be fed a bunch of facts about a news event, then be able to converse with a news consumer and share the facts with each user in a unique and personal way.
Of course, texting back and forth on your phone to learn about a news event might be a bit of a pain. More promising is in the near future when we have much-improved speech recognition combined with artificial intelligence. Imagine driving in your car and simply conversing with its entertainment system (or your phone) in the manner described above; it would be like talking to a human who can tell you what you want to know about a news story.
Top image: A screen grab of the writer’s interface using the Massively authoring tool.