Computer Generated Text Woos Venture Money

The NYT has proof that news about the weather and sports needs no human touch, if all the reader wants is a list of facts loosely strung together

The company’s software takes data, like that from sports statistics, company financial reports and housing starts and sales, and turns it into articles. For years, programmers have experimented with software that wrote such articles, typically for sports events, but these efforts had a formulaic, fill-in-the-blank style. They read as if a machine wrote them.

Computers writing prose and poetry and news have been around for as long as the Internet itself. Have you tried the Elizabethan insult generator or the random prose generator?

The important shift in this story is that people who want to make a high return on their investment are getting interested in pushing to make a sizable profit from the automation of language. The issue thus becomes whether pressure to build this new market will bleed straight into phishers and spammers who have already proven they know how to use automation to turn the news into easy money.

The NYT shamefully gives no credit to attackers who have been successfully composing text from automation for years, nor do they mention the risks from blindly accepting computer-generated text as worthy of consumption.

The few sports writers I read have a specific style and sense of humor. I’m not interested in the data on the game, since I could get that anywhere, but rather how they interpret and present the information based on their particular/unique view of the world.

Update: And in the end a computer that has a particular view of the world will still be just a shallow reflection of the person who programmed it. My presentation at BSidesLV 2011 (2011: A Cloud Odyssey) addressed this issue in great depth. I’ve been asked a lot lately to post the slides from that talk so I’ll upload them soon.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.