I don’t talk about the publishing industry very often on this blog. That might seem a little odd–this is a writer’s blog, after all. But in reality, most publishing news is concerned with royalty rates, rights management, and corporate mergers and acquisitions. Riveting for writers, but sleep-inducing for 99% of my readers.
However, every so often a story with more general appeal appeal turns up. Oddly, they seem to travel in packs. Today, we’ve got a quartet of publishing stories with technological angles.
Over at Publishing Perspectives, Mark Piesing has a piece on robojournalism and robowriting. Fascinating, really. Yahoo uses software to write thousands of stories every week for its Fantasy Sports sites. AP uses programs to generate thousands of quarterly earning reports. Some companies use the technology to produce company reports and online customer service. If you’ve ever thought the customer service rep in your online chat seemed a bit robotic, you might have been right. But perhaps not: Mr. Piesing cites a study in which readers found a robot-written report of a football game more credible than one written by a real reporter.
Mr. Piesing seems to believe that robojournalism is the future of the industry–he tells us that, according to unnamed commentators, “90% of journalism read by the public will be written by robots” within a decade.
My immediate reaction was that he was hopelessly pessimistic. I–and Mr. Piesing’s experts–don’t believe software will ever be able to replace investigative reporters. Can you imagine a program replacing our buddy Jaxon: uncovering and reporting the multitude of human failures that led to the Bay Bridge Bolt Botch?
But then I started thinking more deeply. There’s a key phrase in that quote above: “read by the public”. How many people are actually reading Jaxon’s articles? How many more people read the headline, think “Another boring bridge problem,” and move on to the Sports section?
Next time you go shopping, take a look at the magazines near the register. How many people do you see sneaking a peek inside? The circulation figures don’t begin to give a full picture of how many people read those journals.
Can you honestly tell me that any of the articles were written by humans?
Amazon is getting a fair amount of press for its Kindle Convert software. Stories are billing it as the text equivalent of CD ripping software, in that it allows you to convert your physical books into ebooks (Kindle format only, of course), just as a CD can be converted into mp3s.
Never mind the fact that the user reviews are uniformly negative, or that scanning books is considerably more labor-intensive than ripping a CD. News stories talk about Amazon’s encouragement of piracy and conspiracy to steal from authors. Let’s face it: ebooks are already easily available across the Internet. Kindle Convert isn’t going to make a bit of difference in either direction. The current Amazon sales rank (#41 in Software, squarely between Norton Security and H&R Block Tax Software) is a direct reflection of the current free publicity and heavily discounted price ($19, previously $49). My suspicion is that the overwhelming majority of the copies sold since the price drop will either never be used, or will be used to convert illicitly-downloaded books into a Kindle-friendly format.
Speaking of Amazon, have you considered the implications of their recommendations? When you look at a book, they suggest others you should buy. Those recommendations come, not from any similarity between the books or any assessment of quality, but from purchase records: other people bought them at the same time they bought the one you’re looking at.
A company called trajectory thinks they have a better
mousetraprecommendation tool. According to a piece in Publishers Weekly, their software classifies the content of books with regard to more than thirty attributes, including mood, pace, and intensity. That allows them to recommend similar books in very much the same way that Pandora recommends similar music. And, as with Pandora, the more books they’re able to classify, the better their recommendations will get.
Currently, trajectory’s focus is on selling to libraries and booksellers, but they’re also looking to sell recommendations to self-published authors. That’s a little worrisome. If writers can bias the system to preferentially promote their book when certain attributes come up, what’s the point of trajectory doing the analysis in the first place? And, speaking from the author’s point of view, if trajectory catches on with booksellers, getting that preferential placement could become a “must do”. Can you imagine the bidding wars for keywords that would link to your book from the current bestsellers?
Finally, here’s an art project that completely misses the point. Thijs Biersteker has created a book cover that uses facial recognition software to decide whether to allow you to read the book. If you have a skeptical expression, the cover engages a lock, preventing you from opening the book.
The artist explains that the book has “already been judged and awarded by an international jury, so all that is left is to approach the content of this book without any judgement.”
Excuse me? I should abandon my own critical faculties and make my reading decision solely on the basis of a stranger’s opinion?
I guess that’s one way to avoid bad reviews. If I can’t form my own opinion, I certainly can’t do a proper review.
Hey, maybe I can put a mask showing the proper “serious reading expression” on my scanner, use Kindle Convert to convert Biersteker’s book into electronic form, and then feed it to trajectory’s software, and pass the output to a roboreporter to write my review.
Bets on the number of stars RoboReviewer will give?