Digital news publishers need content, and to produce that content they need reporters. But staff reporters don’t come cheap, making local news publishing a challenging business to be in right now. Could local news automation be the solution that digital publishers have been looking for? The answer depends on who you ask.
For the companies creating the latest class of local news automation solutions, the answer is a firm yes. Technology vendors are leveraging openly available government datasets and artificial intelligence tools to help digital news reporters quickly construct locally focused stories.
The Reporters And Data And Robots project, known as RADAR, is one example of a solution that relies on data-driven templates with fragments of text and ‘if-this-then-that’ rules, designed to translate data into location-specific news stories.
For example, an article written with a RADAR template might begin with a paragraph like this:
“U.S. Department of Transportation records show that [NUMBER] people were killed and [NUMBER] people seriously injured on [CITY]’s roads in 2017.”
The template can easy be adapted based on the location, with the number of accidents and the city name going straight into the text.
RADAR’s in-house team of data journalists brainstorm new angles and storylines for the project, and then puts basic information into each story to give it some local context. In RADAR’s project, local news automation is considered a “production assistant,” quickly adapting basic text in the template to localize individual stories. Reporters and editors working for digital publications can still adapt stories to make them more relevant to local readers.
With a solution like this, just think of the possibilities.
When Jeff Bezos purchased The Washington Post back in 2013, local news automation was still in its infancy. Since that time, his newspaper has become a leader in the space. In 2016, The Washington Post debuted its own news bot — Heliograf. Heliograf is considered to be one of most sophisticated uses of artificial intelligence in news writing to date.
Heliograf can auto-publish stories about basic news events, like Olympics coverage and political election updates, but it does require some handholding from human editors.
Editors using Heliograf must write templates for stories, making sure to include phrases that account for different potential outcomes. Then they connect Heliograf to a reputable data source. (In an election, that might be a data website like VoteSmart.org.) The software identifies relevant data, matches it with the phrases in the template, and publishes the article across multiple platforms. If Heliograf finds any anomalies in the data, the bot will notify reporters, so they can investigate and potentially update the article.
While it might seem like Heliograf requires a lot of handholding, the bot is actually capable of generating a huge number of articles about local topics. For example, Heliograf could churn out articles about the results of every local election race in a state in an instant, as soon as the results are released. Compare that to the length of time it would take a reporter to manually put together the information and publish it to the web, and you can see why local news automation is such a fascinating space for digital publishers.
Heliograf is available to clients of Arc Publishing, which is a digital platform and suite of publishing tools built by The Washington Post. However, it’s not the only local news automation solution available to publishers today.
Let’s look at some other popular options for publishers interested in local news automation.
Wordsmith, by Automated Insights, is a natural language platform that turns data into narratives. The self-service platform allows for total customization of the content it’s creating, along with real-time updates, and an API for larger publishers that need a bit more flexibility. Wordsmith can be setup to create stories out of nearly any type of dataset, including earnings releases.
Using a similar method to produce stories as Wordsmith, Quill relies on datasets and algorithms to generate local articles for newsrooms around the country. For an example of how the platform can be used, check out ProPublica’s investigation into the U.S. Department of Education. ProPublica used automation tools to generate individual narratives for each of the 52,000 schools in the U.S. Department of Education’s database.
MediaCentral | Newsroom Management
Another company selling news automation solutions is Avid, a media technology provider. Avid’s MediaCentral | Newsroom Management platform provides automated content indexing to help reporters find media assets from any location and ultimately turn around stories faster across multiple publishing channels. Avid’s MediaCentral | Newsroom Management platform was designed for newsrooms of all sizes, including local and regional publishers.
Have you considered using local news automation tools at your publication? If you’re interested in learning more about the latest strategies in digital publishing, and how automation fits into the equation, give us a shout.