More than a year ago, The Economist published a story on 3D printing. This is a manufacturing technology which transforms the blueprint on your screen into a solid thing on your desk.
It sounded amazing regardless of the fact that it could take away the employment of factory workers.
But skip ahead into the present day and it’s not so wonderful when there’s the prospect of losing your own job to a computer. Instead of the looming death of print, there’s a new monster on journalism’s horizon:
Narrative Science, a Chicago-based company, created a system that transforms data into stories. The technology combines artificial intelligence and big data analytics to produce articles that read as if written by humans.
Co-Founder and Chief Technology Officer, Kris Hammond, says, “We literally go from numbers and symbols to language… in sports, finance, real estate, politics, marketing data.”
"FINAL: William Buford scored a career-high 29 points and grabbed seven rebounds to lift No. 3 Ohio State to an 87-84 win over Purdue on Tuesday at Value City Arena. Buford hit a big field goal with 4:30 remaining, giving Ohio State (21-3) the crunch time lead. Defense was an afterthought as both teams shot 52 percent."
The system ‘understands’ the stories that it produces. It can formulate the same ideas in various formats such as PDF files, business reports, and Tweets. It’s even able to discern the most valuable aspects of a text.
Companies like Hanley Wood have already used the program for their site builderonline.com. Others such as Forbes magazine and the Big Ten Network also use it.
According to The New York Times, Andrew Reid from Hanley Wood says he’s impressed not only by the quality of the articles, but by the cost as well. $10 per 500-word article is cheaper than hiring actual people to write.
The Terminator takes over
In a 2011 Strata Summit interview, Hammond said machine-generated content is based upon data. However, he notes that it’s also associated with the notion of narrative structure.
One can fill in that narrative structure through search and crowdsourcing, and so the technology has the potential to expand in the creative realm too. Hammond says he doubts they’ll take it that far, but who says that others won’t?
Of course, computers can’t write about emotions. They’re not creative or original, and cannot express art forms.
Not yet anyway.
Soon this technology might be capable of producing profiles, features, and real investigative journalism. Perhaps it will even write a novel.
And what about the media industry? Forget print or advertising blues; it’s the human journalists that are in trouble. Old fears will evaporate if this technology becomes the future of journalism.
If you can’t beat them…
This is a lesson for media workers. The danger of being replaced by a machine might motivate journalists to replace generic content with first-rate articles (which used to be the priority).
The technology could relieve writers. Instead of wading through a sea of data, one could use this tool to curate the most important facts for a story.
Hammond views the program’s stories as core content, which journalists can expand on. In this way it’s not very different from using a newswire service.
Narrative Science is attempting to cover The Long Tail of news and information. Not only are they generating content for areas where there’s no coverage, but also content that people aren’t capable of tracking.
For example market summaries for real estate, where people don’t have logistical power or staffing to write those stories. Also, the technology builds stock alerts for clients by examining the stock market every few minutes. This is not practical for normal journalists, but very useful for the clients.
This new system poses both threats and benefits; it can help journalists and organisations, but it can also take away jobs.
One thing is for sure: technology exceeds its limitations every moment.
The days of humans are numbered.