That’s the question enlivening discussion on a Linked In group for journalists in mid-November. Someone posted to the group an article by Julie Segovia from the website Ink’d titled, Invasion of the Writer Snatchers. The article asked the question “Can stories produced by modern-day robots stand in lieu of those written by humans?”
It described the software platform Quill, designed by a Chicago-based company called Narrative Science Inc. The software has the potential to automatically generate stories that hinge on data, such as baseball games, real estate, financial services, marketing services, and intelligence (the industry, not the quality).
Segovia’s article states that the health care, retail, and pharmaceutical industries have indicated interest in the software, and that the reason for such interest is obvious: robots, or bots, as Segovia calls them, are cheaper than humans. They won’t demand raises, health insurance or a retirement plan. They will never take vacations or request time off to tend to a sick child.
The Narrative Science website states:
Quill applies complex and sophisticated artificial intelligence algorithms that extract the key facts and interesting insights from the data and transforms them into stories. The resulting content is as good or better than your best analyst, and is produced at a scale and speed only possible with technology – technology that is now patented.
In a sense, this is not news to me. I was aware many years ago of so-called “canned” astrology reports, built on a person’s birth data. Astrologers use these reports to provide basic information quickly and cheaply, in the hopes of getting the person who asked for the report to come in for a longer, more personalized (and more expensive) chart reading. What Quill is doing is a more sophisticated application of the astrology report writer.
One person in our Linked In group said that he could see Quill making stories out of baseball box scores, but that he could not see a robot reporter covering a local school board meeting. That is where I jumped in to disagree. I have covered a lot of school board meetings, and, now that government meetings are audio and video recorded for broadcast, I foresee the day when a robot that has been given the agenda and the recordings writes a story. We only await voice recognition software, now a standard feature on new PCs, to accurately transcribe multiple voices rather than one it is trained to understand now. That day is coming because journalists want that capability, so programmers are working on it.
In the end, computer-generated news stories will be accepted if the public, or a targeted industry readership wants them. So the questions are: what does the readership want and why? Do people care about whether the source of their news is human or mechanical?
One member of our group said that he did not see software programs such as Quill ever creating stories of a type deeper than syndicated AP wire copy. AP wire copy is vital to short-staffed community and college radio news departments that depend on volunteers. One can criticize the journalistic depth of most “rip and read” stories, but humans are now writing that copy. Replacing the writers with software will not raise the quality; It will just eliminate jobs.
If the software can do no better than that—they may do better in 10, 20, or 50 years—and software widely replaces human reporters for cost control reasons, will people miss human journalists? If people can be deliberately trained by corporations, or ingrained from habit, to expect and want only the quality of writing they are given, and if future generations read only a level of writing that would have been considered substandard in the past, will human journalists even have a chance in a world driven by the corporate bottom line?
The “big data”-based story is the type most easily handled mechanically, and we know that these days, “big data” is all the rage. But human reporters are being trained to deal with “big data”. The organization Investigative Reporters and Editors has been holding workshops and boot camps on CAR – Computer Aided Reporting— since long before Quill was invented. There are some things that computers do better than humans, but writing and analyzing data is not necessarily one of them, at least at the moment.
Still, according to Segovia’s article, Kris Hammond, a co-founder of Narrative Science Inc., doesn’t believe readers will have to miss even high quality writing when robots replace human writers. He has boldly predicted that “[i]n five years, a computer program will win a Pulitzer Prize — and I’ll be damned if it’s not our technology.”
I believe that is wishful thinking on Hammond’s part because I do not believe machines can acquire the imagination and emotional depth of a quality human writer. Of course, I may be indulging in a bit of wishful thinking myself. The quality of human writing is declining and has been, long before PCs and the Internet came along. It is declining with the general decline in the quality of education.
One of our members, David Rosman, writes editorials for the Columbia Missourian and book reviews for the New York Journal of Books. He said, “The art of writing has evolved. I am reading a new book, ‘The letters of John F. Kennedy’ by Martin Sandler, and am becoming very aware of how our language and styles have changed. One of the reasons I personally enjoy reading the “classics” is for the language.
In journalism, long-form reporting is becoming a lost art. We are hit with 140 characters from CNN, Washington Post and the AP with little, if any, follow-up. We are losing magazines and the analysis of political, economic and social issues. Even the Onion has lost some of its charm by going 100-percent digital. We are getting stories in bits and pieces, much of which seems unconnected or a rehash of old news. As the cycle becomes shorter, as events in Syria are reported with the same emphasis as a shooting in Seattle with little follow-up, we lose sight of the meaning of both.
People are hungry for information and entertainment, something the written word does better than other media. Writing will not die, but continue to evolve to meet the wants of an audience that is moving too fast for its own good.”
One can blame technology for shaping the unfortunate changes in writing that we have seen lately. But technology facilitates the desires of its users. The 140 character limit of tweets and texts made us write Bingo instead of English; “Before” is now B4″. But look to humans, not technology, to see who is to blame for a world full of fast pace and tiny space. The “audience that is moving too fast for its own good” crams its life with much useless information, even as it develops a technology-induced ADD that makes it prey to sound bite politics instead of investigative journalism.
I am a Baby Boomer. Computers came into my life in middle age, long after my habits were supposedly set. Yet I mutter, “Oh radio interviewee, why did your book have to be over 400 pages long?” Yes, the Internet does change your brain. But my choice to spend 8 or more hours a day on a computer facilitates that change.
What happens to the generations that follow mine? Will they still want to read the classics for the language? Or will they take a “Just the facts, Ma’am” approach, not caring if the purveyor of those facts is man or machine?
We use artificial intelligence (AI) to make machines more like people, while at the same time, employers, especially, demand people to be more like machines—to multitask, to not have human needs to eat or go to the bathroom while at work, to not want vacations or sick leave. Why can’t people be people and machines be machines, each doing what they do best?
Assisting humans in performing precision work, such as surgery, and replacing humans when it is too dangerous (Fukushima’s crippled reactors) or otherwise not feasible for them to be physically present (exploring Mars) is the proper purview of robotics.
Replacing humans in tasks that people can capably do, such as writing, raises questions about what we are supposed to do, especially in a money and jobs-based society in which we are expected to “earn a living”. But even in a fully demonetized society, the question would remain if robots did it all for us. With governments looking to robotize war, we might not even be fit for cannon fodder in 100 years. Will we become the obese lounger-bound consumers of the movie Wall-E?
Will people become obsolete?