How Journalists Can Use AI: Add Context Based on Archives

In an era when misinformation can be rampant around the web, it’s so important to have trusted sources you can turn to for accurate news. One of the most credible sources, that you know you can trust, is quite simply your own team.

John Byrne, founder and CEO of Raw Story, approached us with an interesting idea. He wanted to build an AI editorial tool from the ground up that was based on their content archives. The independent news website, which recently turned 20 years old, is based on delivering news and insights at the breakneck speed of today’s political news cycle. It’s endless. Time is of the essence, as new content is pushed out constantly, so becoming more efficient is really important to their team.

Enter the AI chatbot that our team built in partnership with them. Trained on more than 300,000 posts in their archives, it has encyclopedic knowledge of political events, important people, key organizations, and mainstay topics over the past two decades. It has the ability to write up high-quality background on issues that Raw Story has previously covered, all in a fraction of the time it would take someone to do the research.

The chatbot in action, linking to the most relevant stories about a specific person.

“It’s best at generating more background copy, and historical information, that we can put into breaking news to make it longer, more cohesive, and more understandable for our readers,” said Raw Story Executive Editor Adam Nichols.

It’s not only based on the text of Raw Story’s articles, it’s much more intelligent than that. It has authors completely indexed, so you can search for content written by a specific journalist. It has publish dates for every article, so you can set timeline filters to find the exact context you need. It provides links to stories in Raw Story’s archive, so it’s a copy and a paste away for better linking of all of their content together in a smart way. And, of course, this is Raw Story — it builds upon itself nearly in real time. You can search for relevant content from today, or this week.

It is conversation-based and extremely intuitive, so it knows the ways it can best help an editor or writer on the fly.

• Need to add background to a breaking news story? Easy.
• Looking for context on that court ruling today? Simple.
• Trying to build a timeline of events for a complex issue or topic? Done.
• Seeking links to improve the reader experience? Sure.
• Want to proofread an article and make suggestions for improvements? All set.

Has it been useful in accomplishing the goals they set out for it? “Unquestionably,” said Nichols. “It is accurate and it is very easy to edit down.”

The chatbot has a sense of time so that the most relevant links and background information are surfaced.

Between its assistance with copyediting and adding context, Nichols estimates that the tool cuts out a solid five-to-six minutes for Raw Story’s editorial process, which in the minute-by-minute political world in which they race to beat competitors, is extremely significant. “Quite substantial,” he said.

Byrne said that one use case he’s seen is utilizing the bot instead of Google Search to quickly find content Raw Story has previously written. It identifies the right links and context for newer stories.

“It does a better job of taking all the articles together, condensing it to a short thing, and giving you the background,” he said.

It’s a great editor, too.

“It helps speed up the editing process by correcting grammar in a more sophisticated way than tools have before,” Byrne added.

All in all, it makes the team at Raw Story much more efficient.

If you’re interested in building an AI-powered chatbot like it for your own editorial team, get in touch with us! We can help you with exactly this kind of solution to turbocharge your team and improve efficiency.