Careful readers of my SDTimes posts – and surely, dear Reader, I may count you among their number – have no doubt noticed that despite the failure of my faith in AI's promise, I remain interested in the technology. Or technologies, I suppose I should say. It's fascinating stuff, always at the cutting edge of computer science.
In fact, it is the cutting-edge nature of artificial intelligence technologies that makes them interesting. AI projects always seem to incorporate technologies that aren't quite ready for widespread adoption. Once the technologies migrate out of the lab and into the world, they're not quite so interesting anymore. They quickly become mundane elements of everyday life. The AI magic is gone.
The grammar checker in your word-processing application is a case in point. It wasn't so long ago that highlighting grammatical errors in written work was considered beyond the reach of ordinary algorithms. Grammar checking was strictly AI. But once the nut was cracked, the ability to analyze sentences, highlight errors, and volunteer remediation became an ordinary part of every word-processing application. Once it was magic. Now it's a baseline requirement. You can't even charge extra for it.
As an AI aficionado, I suppose I ought to celebrate when a technology finds real-world application outside the labs. And I do, honestly. But I mourn a little too. Because a bit of magic leaves the world whenever an AI technology goes mainstream.
You understand, therefore, my mixed reaction to the revelation that an AI program is now writing news stories – actual articles – for publication.
This is no demo program working in a laboratory. It's a real live application that is even now being used by publishers to cut writers out of the loop and reduce the cost of generating news stories. (Do I sound strident? I don't mean to. But I'm one of those writers, you know. It's only natural that I might feel a bit threatened.)
The company behind the technology is Narrative Science Inc. The founders' credentials include experience at high-tech start-ups and Ivy League universities.
The customer list is confidential, but a handful of publishers have disclosed their use of the technology. Big Ten Network, for example, uses a Narrative Science application to prepare play-by-play game summaries for its Web site. The stories appear on the BTN Web site within minutes after the game ends. Here's an example: Final: Wisconsin 51, UNLV 17.
Hanley Wood, a publisher of newsletters and Web sites in the construction industry, is another customer. The company uses a Narrative Science application to write stories based on monthly business statistics. Here's an example: New York, NY – Local Housing Data – Summary.
These examples are impressive. The stories aren't creatively written, but they aren't bad. They're useful. Readable. Publishable.
The best news, I suppose, is that Narrative Science is hiring. If you're willing to work in the Chicago area, check the Web site for current openings. It must be a lot of fun to work on smart apps like these.
Web recommendation: When Wired was launched as a print magazine – way back in 1993 – it was the most exciting magazine I had ever seen, packed with well-written articles on breaking technologies. Yes, it was geared to consumers, not developers. But it was thrilling to receive a monthly overview of what was cool in high-tech. It's online now, of course, and still worth reading. You can find it here: wired.com. J.D. says check it out.
J.D. Hildebrand has written hundreds of articles for dozens of publications and online communities dedicated to software development. He recently relocated to a small town outside Belgrade – stop by if your travels take you through Serbia.