What it looks like when the wheels fall off; Or, Where’s the good-journalism bot?

The software tools mentioned in These are the bots powering Jeff Bezos’ Washington Post efforts to build a modern digital newspaper:

Feels Bot

Asks you to express your feelings about a news topic by selecting from a pre-chosen list of five emoji. Hours later, shows you how well your take on the topic conformed to what it claims are the opinions of others. Important stuff, right? This is a good bot because it gets a lot of engagement. A public-facing bot on Facebook Messenger.

Virality Oracle

Half an hour after a WaPo story goes live, this bot begins issuing forecasts, accessible to WaPo staff, of how much chatter the story is going to generate, according to a really amazing algorithm from our data science team. Based on these predictions, a story can be tweaked to make it more buzz-worthy or, if it’s doing well, used as a template for other stories: we can maybe go in and add something of the story, or start writing off of it a little bit. A Slack bot on an internal WaPo Slack channel.

MartyBot

A bot that reminds reporters of upcoming deadlines for in-progress stories.

Heliograf

Anybody every play Mad Libs? You know, put in a noun, pick an adjective, whatever? This is kind of what that does. Why is this a good thing? We published a story on every single Olympic event, because of Heliograf. And then for elections, we posted a story on every single race in the U.S. on Election Night, and generated newsletters, generated tweets. If the analogy to Mad Libs doesn’t grab you, you can always think of this bot as being a modern-day incarnation of the novel-writing machines from Nineteen Eighty-Four. Wonderful, no?

The NiemanLab post linked above is an edited transcript of a talk delivered at an online journalism confab by WaPo‘s “director of product”, Joey Marburger. Marburger began his presentation with a tongue-in-cheek set of three laws for journalism bots, referencing Asimov’s Three Laws of Robotics. His zeroth law states, A journalism bot may not misinform a reader through out-of-contest information, express opinion on non-opinion hard news, or ignore corrected information when updated.

How is that relevant to any of the bots subsequently described by Marburger? One of them nags reporters about their deadlines. Another churns out extruded news product that allows the WaPo to claim it has published something about a given topic every single day over a certain period. That probably helps with engagement. The others are about increasing engagement.

Which bot has anything whatsoever to do with improving the quality of the journalism coming out of The Washington Post?

Some of you may be wishing you’d lived your life a bit differently, so you would have amassed a large enough pile-o-money to purchase the Washington Post before Bezos got around to doing it and turning it into the kind of organization that has a “director of product”. That feeling will probably pass.