Recently, Reuters has been creating an AI tool that intends to assist journalists in analyzing data, writing some sentences, and suggesting story ideas.
The tool aims at achieving all these functions without having to phase out reporters, but instead, provide them with a copywriting assistant-cum-digital data scientist.
Since summer the tool, dubbed Lynx Insight, has been tried by numerous journalists, and will now be unveiled across all Reuters’ newsrooms.
According to Reuters’ executive editor of editorial operations, data and innovations, Reg Chua, Lynx Insight aims at allowing machines to do what they do best in a bid to help the human editorial staff succeed in their work.
He pointed out that the recent efforts differ significantly from past editorial tech activities that aimed at training AI to write complete stories.
Reuters’ latest AI tool is expected to go through enormous datasets in an attempt to search for anything fascinating.
In turn, journalists are handed the information coupled with important background and context via messenger service, their data terminals or email. These details help them to kick-start their research when they feel the urge to pursue a given story.
Although Reuters has made an incredible achievement venturing into AI, it is not the first newsgroup to do so.
In fact, The Washington Post has executed hundreds of machine-written brief snippets using its Heliograf, an in-house robot reporter.
MORE: 10 Amazing Examples Of Natural Language Processing
On the other hand, the Press Association is collaborating with Urbs Media on the RADAR project for the local news’ natural-language production. Also, Yahoo! has been using Wordsmith, an automated writing tool, for a variety of its material including Game of Throne recaps and football snippets.
Although Lynx Insight can construct sentences, its main potential revolves around data supply.
Thanks to this tool, journalists do not have to trawl through voluminous databases and skim numerous reports to find details for their stories. According to Padraic Cassidy, machine learning will instead give them the desired help.
Since not all data points may be useful for journalists, they can identify less insightful parts and inform developers. Doing so not only allows the system to learn but also find better stories.
This function strives to get rid of the UI layer in an attempt to make the system generate facts, contextual background information, and facts while the journalist types.
Even though Lynx Insight currently works with Reuters’ enormous amount of financial data, it may soon be used in sports as well as the organizations’ vast legal databases.
The tool also works on photos and videos thanks to metadata. In addition, Chua believes that Lynx Insight could be utilized in personalizing new stories.
David Caswell, the founder of Structured Stories news platform, holds the belief that AI could help relieve the pressure in newsrooms as journalism revenues drop.
This case applies mostly to news segments that do not generate a lot of money like local reporting. With that said, Chua hopes that the Lynx Insight tool becomes part of a journalist’s day whereby it works in the background like Google or a telephone.
Source Wired
