Robo-journalism: How afraid should we be?
- Hein Marais
- Life, etc
- 06 Jun 2012 01:25 (South Africa)
Much of the news we get these days isn’t news. It certainly isn’t investigative news. In fact, it’s opinion, press releases and, increasingly, algorithms. The scary part is that the new way of interpreting data was partially designed by a journalism school. By HEIN MARAIS.
Imagine discovering that a computer algorithm writes the news you read. Not “selects”—that’s already happening—but actually writes it. Let’s skip the “imagine” part and go straight to Forbes magazine for a taste. The byline credits the story to Narrative Science, and at the tail of the report, there’s this notice: “Narrative Science, through its proprietary artificial intelligence platform, transforms data into stories and insights.”
In this case, the “proprietary artificial intelligence” part is another way of saying algorithm.
Narrative Science is a US start-up that has designed software capable of generating news reports on various topics. Using algorithms, it collates masses of data on the desired topic, selects pertinent information, and narrates it into a simple article or report.
Here’s their pitch: The service “creates rich narrative content from data. Narratives are seamlessly created from structured data sources and can be fully customized to fit a customer’s voice, style and tone. Stories are created in multiple formats, including long form stories, headlines, Tweets and industry reports with graphical visualizations. Multiple versions of the same story can be created to customize the content for each audience’s specific needs.”
Wired magazine’s Stephen Levy has written up an informative gosh-wow piece that explains the 1s and 0s of the process.
Perhaps the saddest detail is the fact that the technology was developed in part at Northwestern’s Medill School of Journalism, Media and Integrated Marketing Communications. Talk about turkeys voting for Xmas.
The company claims that the technology “frees up existing editorial resources to focus on stories that either require deeper investigation or create stories that would not otherwise be written”.
That’s very unlikely. This isn’t about saving journalism - it’s one more way to trim costs and amp up profits.
The economics are obvious. It’s cheaper to subscribe to the algorithm-spawn than it is to pay a pod of young reporters a living wage to crank out such customised products. Each 500-word article costs less than $10, and that price is expected to drop.
Specialist news services are subscribing to Narrative Science for niche sports, investment and other reports. This year, for example, Narrative Science expects to generate 1.5 million accounts of Little League baseball games in the USA.
OK, so it’s not assembling the front page of the New York Times. Yet. But it evokes the spectre of robojournalism—and that offends and scares many in the trade and beyond. Should it, really?
Already a great deal of the lifestyle, travel, entertainment, sports and business pages of newspapers bear only marginal evidence of independent, journalistic intervention. Recycled media releases constitute increasing proportions of that “soft” content.
Worse, as newsrooms are gutted, multiple “hard” beats are being stacked on the desks of bushy-tailed newcomers who, for all their wound-up enthusiasm, are mostly and frantically treading water.
At the courts for a crime story in the AM, then off to a press conference on a new public housing scheme, then another press conference as a trade union weighs in on changes to health and safety regulations. All the while, filing updates for the web on each.
There’s not much journalism happening on a day like that, very little accumulated savvy or newshound scepticism being deployed, hardly any investigation in evidence.
And the journalist is thrashing away at it all for less money, with fewer benefits, hardly any authority and meagre experience. That’s at one end. At the other is the bubbly effluent of “opinion” and the branded columnists whose task is not to share insight and understanding, but to titillate, amuse and antagonise. Nothing intrinsically wrong with that, but it’s not journalism.
Sandwiched in between is an embattled anachronism: the journalist who knows her beat, who has the time and license to research, to dig, to help cast some light on the world, and to embarrass corporate and political power when that’s merited by the facts. Prizes still exist to honour the craft of those artisans, but with each passing year the laurels look more like mementoes of a bygone time.
This wobbly amalgam is the traditional news business “streamlined” (read: eviscerated) by shareholder interests, squeezed by the internet, pushing product, mainlining “efficiency”, chasing profit. Enter: the algorithms.
Narrative Science’s chief technology officer reckons more than 90% of news will be written by computers in 15 years’ time. In a way, that forecast seems out-of-date.
Large volumes of “news” already are written either (via media releases) by the subjects of the story or by harried, dynamic but befuddled drones. The line separating algorithmic news from the current state of things is fuzzier than we realise. DM
- Hein Marais' blog can be viewed here. His latest book is South Africa Pushed to the Limit: The political economy of change (UCT Press & Zed Books, 2011).