Contrary to popular belief, the steadfast march toward automation is affecting all sorts of fields — not strictly blue-collar industries like manufacturing and transportation. Already, artificially intelligent systems (AI) are reviewing contracts and mining documents in discovery; determining which job candidates get callbacks; and selecting the inventory retailers choose to highlight for particular customers.

And now, at least one publication’s using it to pen first drafts of stories.

According to a report this morning in Digiday, Forbes recently rolled out a new tool in its content management system (CMS) — Bertie — that composes rough article drafts which merely need to be fine-tuned rather than written from scratch. It’s an evolution of the business publisher’s semiautomated topic recommendation feature, which surfaces news topics based on writers’ previous work, and it’s expected to become available to all Forbes contributors in North America and Europe early this year.

Salah Zalatimo, Forbes Media’s newly appointed chief digital officer, described the robot-penned pieces as “thought starters” rather than publishable work. Toward that end, the tool sources both Forbes and competitors for links to contextually relevant articles about topics, along with images that might improve the story.

It’s not unlike Reuters Lynx Insights tool in that respect, which launched in March. Like Bertie, Lynx Insights surfaces key data related to stories — helping reporters to, for example, quickly analyze historical trends in commodities pricing.

And it reflects something of a trend. As Digiday notes, Forbes and Reuters are far from the only news publishers experimenting with AI reporting tools.

The Washington Post’s in-house Heliograf platform, which generates short stories on a range of topics like the Olympics, congressional and gubernatorial races, and high school football games, spit out 850 articles in 2017, a number which grew to “thousands” in 2018. The Associated Press, meanwhile, in partnership with startup Automated Insights, deployed an AI writer in 2015 that’s able to produce roughly 2,000 articles a second with fewer errors than their human-produced equivalents.

Despite how it might seem, AI-assisted reporting platforms aren’t necessarily a harbinger of machine-driven newsrooms, Jeremy Gilbert, director of strategic initiatives at the Washington Post, contends.

In an interview with Digiday, he says that tools like Heliograf can spot unexpected trends in news events, or undertake some of the time-consuming legwork currently performed by human reporters, like identifying trends in vast financial datasets. In 2014, for example, the LA Times used a machine learning algorithm to comb through eight years’ worth of public records — findings that contributed to a report on the Los Angeles Police Department’s history of misclassifying violent crimes.

“We think we can help people find interesting stories,” he said.

That jibes with the AP’s strategy. It estimates that its automation tools have freed up 20 percent of reporters’ time spent covering corporate earnings alone.

“One of the things we really wanted reporters to be able to do was when earnings came out to not have to focus on the initial numbers,” Philana Patterson, an assistant business editor at the AP, told The Verge in an interview. “That’s the goal, to write smarter pieces and more interesting stories.”

Indeed, the Tow Center predicted in a 2016 report that automated journalism “will likely replace journalists who merely cover routine topics.” And in a survey published by Tata Communications in September, Ken Goldberg, a leading AI researcher and UC Berkeley professor, said that he expects AI’s continued advance into workplaces (including newsrooms) won’t come at the expense of jobs, but rather will “create new ways of working” and “new jobs” in companies.

“Robots and AI are not going to take away this creative, insightful, empathetic aspect of almost every job,” Goldberg said.