Editorial
Robots Are Coming ...
TO WRITE YOUR STORIES. Every task for which an algorithm can be created can now be done better by computers than humans. That is, if you can describe in detail a particular task, then a computer program can be written (and a computerized machine constructed) that will enable that task to be done better than any human could possibly do it. The problem is that many tasks are exceedingly difficult to describe exactly; and many tasks, even if well-described, remain at risk of requiring unforeseen changes if something suddenly goes off the rails. Imagine a heart-surgery robot, or even an algorithm for heart surgery.
On the other hand, a computer will never get distracted, drunk, or tired when driving a vehicle. It might hit a few orange cones that a human could possibly avoid, but on balance, it could potentially do a much better job.
What about those special talents that human beings consider “being creative?” Could they be done by computer? Maybe not. In a general sense, the computer has no human experience from which to draw. There are attempts to remedy this by encoding databases full of the descriptions of common human relationships to objects and events. But even so, describing in detail the human “aesthetic” is a daunting task, and might prove to be impossible.
For some time there have been robot writers. Here is one from Stanford called the Regency Romance Writer which, of course, writes not-very-good romance fiction. The stories continue nearly forever if one continues to refresh the page.
Incense and Prejudice
Waiting alone in the Red Chamber, remembering its bloody history, with the plaintive whinnying of the horses wafting in from outside, Serena thought once more of headstrong Henry Crawford, the hooded visitor in her recurring dream. He was now, according to the gypsy woman, lost forever at sea.
Then came an inrush of cool air as the door was flung open, and she dropped the brimming wine glass heedlessly on the rug. He was here! “Without you I am nothing, Adored One!” he blurted, lifting his quizzing-glass and gazing at her passionately, and as he lovingly clasped her to his manly chest, she started to think about what they would call their children.
One hopes—Robby, Data, Unicron, and if it’s a girl, Dot Matrix or False Maria. Even as strange as human romances are, it is easy to see that the story immediately begins to fall apart. It is also easy to see that it could be done better with a few decades more work. This one appears to be random selection from a menu of events and responses. But imagine how complex this sort of robot writer could be.
Newswriting is being done by computer more and more these days. Here is a report entirely written by a computer algorithm called QuakeBot, programmed by Ken Schwencke, a journalist and programmer for the “Los Angeles Times,” and published by them just minutes after the event:
A shallow magnitude 4.7 earthquake was reported Monday morning five miles from Westwood, California, according to the U.S. Geological Survey. The temblor occurred at 6:25 a.m. Pacific time at a depth of 5.0 miles.
According to the USGS, the epicenter was six miles from Beverly Hills, California, seven miles from Universal City, California, seven miles from Santa Monica, California and 348 miles from Sacramento, California. In the past ten days, there have been no earthquakes magnitude 3.0 and greater centered nearby.
Of course, not much intelligence, artificial or otherwise is required here. There is a simple template that is filled in with data supplied by the USGS. The item is reviewed before publication to eliminate any really grievous goof-ups. That’s all.
I have watched sports “wrap-ups” that could have been done at least as well by machine. They mainly consist of the name of a sports teams (any sport) and the words “beat, trounced, hammered, clobbered, shut-out, overcame, defeated, crushed, annihilated, lost to, rebuffed, put down, over, overthrew, massacred, whipped, knocked out, KO’d, licked, shellacked, pummeled, trampled, battered ...” added in somewhat random order or nuanced—depending on the score—followed by the name of the second team.
I’ll bet TV and radio news readers get exhausted by the search for verbs and the monotonous repetition. Better a machine should do it. I can see an invisible cartoon thought-balloon rising over their heads saying, “I went to college for this? Jeeeeezzzz ...”
Sportswriting is particularly vulnerable to being done by robots because so much of what happens is contained in numbers, rosters, and records. (And there is hell to pay for a human reporter making a reporting error.) Even those post-game locker-room interviews are dreadfully robotic. Recently, “The Washington Post” seriously considered using robot writers to cover the region’s many high-school sports contests. The reporting on the recent Olympics too, was aided by robot reportage, especially in the use of performance charts.
Automated Insights—an up-and-coming company that primarily specializes in publishing sports stories, and even fantasy football stories and summaries— seems to be in a good position to expand into other fields. AI, Inc., was recently awarded U.S. patent 8515737 for Systems for dynamically generating and presenting narrative content.
Narrative Science, Inc., is another company that competes in turning data into human-readable narratives. They lean more towards business and financial content.
You might be surprised to learn that templates and algorithms have been designed to publish complete nonfiction books. In fact, the world’s most published “author” is a gentleman named Philip M. Parker, Professor of Management Science at INSEAD, Fontainebleau, France, who has something like 106,575 different books currently on Amazon, written with substantial assistance from sophisticated computer programs. He claims he has written over one million books, many are now out of print. Examples from Amazon:
“Webster's English to Cornish Crossword Puzzles: Level 1.”
“Webster's Persian (Farsi) to English Crossword Puzzles: Level 1.”
“Webster's English to Croatian Crossword Puzzles: Level 1.”
“Webster's English to Czech Crossword Puzzles: Level 1.”
“Webster's English to Portuguese Crossword Puzzles: Level 27.”
“Webster's Italian to English Crossword Puzzles: Level 9 (Italian Edition).”
You can see how there can be an enormous number of these in the matrix of possible books. Not only are there many languages, but they go both ways in translation. Furthermore, the instructions and miscellaneous text can be in any other language. There are also an uncountable number of levels of difficulty. So far, Level 27 seems to be the top. But the mind boggles at the possibilities.
Professor Parker puts his name on books on many other subjects as well. A sampling of other works:
“Noonan Syndrome—A Bibliography and Dictionary for Physicians, Patients, and Genome Researchers.”
“The 2007-2012 Outlook for Food-Grade Dry Whole Milk Shipped in Bulk in Greater China.”
“Webster's Mohawk—English Thesaurus Dictionary.”
“Marx: Webster's Quotations, Facts and Phrases.”
And his fill-in-the-blanks prolific story generator front end demo is well worth watching. The professor also has a patent on robot-writing (Method and apparatus for automated authoring and marketing, U.S. patent 7266767) and likely intends to sell programs that will enable anyone to be a published author of fiction.
Lambert M. Surhone is a lazy scribbler in comparison with only about 5,000 different books on Amazon (but around 10,000 in his catalog). Admittedly, many are simply semi-robotic republishings of articles from Wikipedia and the links therefrom through his Alphascript and Betascript Publishing Cos. This has not been without controversy. Amazon announced several years ago that they were going to hand him his hat, but he is still listed.
Surhone’s works are more idiosyncratic than Parker’s, but then again he doesn’t seem to employ the fecund matrices of Professor Parker. Surhone’s books are wildly expensive, and nobody seems to read them. For example: “South Carolina Highway 245” (Paperback)—$404.92.
So the robot writing of decently good fiction might be further off than fully robotic heart surgery. Computers don’t yet understand human sensibilities sufficiently to create reasonably good works of fiction. But it is a fool’s game to bet against electronic chips and lines of code.
So when will a computer-written novel hit the shelves? Hah! You’re too late.
Russian Alexander Prokopovich of Astrel-SPb has just published the first book written in three days by a computer program: “True Love.”
An excerpt published by “The St. Petersburg Times” reads:
“Kitty couldn’t fall asleep for a long time. Her nerves were strained as two tight strings, and even a glass of hot wine, that Vronsky made her drink, did not help her. Lying in bed she kept going over and over that monstrous scene at the meadow.”
What? Romance again? What is there about romance fiction that is so attractive to robots? Should we be worried?
So what does the future hold? Even now, robot writing is getting complaints. But nobody can keep the cat in the bag. If you can buy a robot-generated book, in a few years you will be able to buy the robot “author.” There will soon be computer programs for sale that will enable one to write any sort of fiction. Do we still call these button pushers “authors?”
Soon you will be able to load into your Kindle a program like Philip M. Parker’s fiction generator that will write fiction for you (and is modifiable on the fly) based on what you enter into the boxes on the first page, and what it learns about your personal desires, perhaps by watching your pupillary response, reading your heartbeat or galvanic skin response.
Alien Spaceships: No[ ] Yes[ ] How Many?[ ]
Sex: Kinky[ ] Hot[ ] Mild[ ] None[ ]
Post-apocalyptic Earth: No[ ] Yes[ ]
Environmental Disaster: No[ ] Yes[ ]
Number of pages: [ ]
Hero’s Name: ________________________________
Enter.
Push to begin reading.
In conclusion, we note that our esteemed editor, Sam Bellotto Jr., is also a professional crossword puzzle constructor with plenty of his own books available on Amazon. His output might be significantly less than Phil Parker’s but he makes up for it in quality.
Eric M. Jones