The platform, owned by Automated Insights, provides “auto generated, data-heavy articles” on topics such as quarterly earnings and college sports. A beta version is available now on Wordsmith’s website, with a full launch expected in January 2016. The technology is already used by companies such as the Associated Press and Yahoo.
Wordsmith works by creating “branching paths” — conditionally adding words, phrases and sections that can be added, modified and removed depending on the article. Users enter data — such as quarterly figures or sports results — around which these branches are built. This story structure, once created, can be used as a template for an unlimited number of articles. All users have to do is enter new data to create a unique story.
But the technology hasn’t been without its critics. NPR ran a feature in which they pitted their White House correspondent against Wordsmith — while technology beat human on speed, the human written article was, according to NPR listeners who voted online, richer and more engaging. Some argue a robot journalist could never replicate the style of a human journalist in terms of style or insight. But Automated Insights argues that Wordsmith can be stylistically modified so stories can be written as if “a person wrote each one of them individually.”
“We focus on personalised content,” said Automated Insights CEO Robbie Allen in a press release. “Instead of writing one story and hoping a million people read it, Wordsmith can create a million stories targeted at each individual user and their preferences. It’s a story that is totally unique to each user because it is powered by their data.”
Wordsmith isn’t just limited to articles, though — it can also produce client reports, financial summaries and product descriptions. Read more…