Google A.I. Uses Novels to Improve Conversation

The team at Google Brain has devised a way to use books to help computers talk. By feeding an artificial intelligence the text from about 11,000 fiction novels, the team has expanded the A.I.’s ability to have natural human conversations. Essentially, the computer reads these books and becomes a better talker. The variety of expression to be found in these novels is greater when compared to non-fiction books. Encountering a variety of phrases and words that express similar ideas helps computers to understand “the nuance of language”, so that when apps are tasked with communicating with the user, they can do so with more organic speech patterns.

The authors of these novels, however, were never asked permission, let alone compensated for the role of their work in helping to develop these very profitable technologies. The books were taken from what is described as  a Book Corpus comprised of free books by as of yet unpublished authors. Many of these books, however, although free to download are also protected under copyright laws that ensure the information be used as reading material. Whether or not that excludes machine/ for-profit readership is a matter of debate, but Google’s callous decision to use the entirety of the Corpus without so much as contacting the authors has left a sour taste in many of their mouth. Author of Hostile Witness, Rebecca Forster, has this to say of her book being used in the project:

“I take great pride in my craft, and perhaps it was chosen because of that. Which would be great. Or perhaps it was chosen because it was there, because it was free?”

Her book is one of the many that includes a copyright declaration that states it is to be used for  personal enjoyment exclusively. Although the company’s efforts may not seem malignant, it is certainly within Google’s power to at the very least inform the author’s of what their work was going to be used for. A matter of principle that would mean a lot, especially given the tech giant’s previous disagreements with the Author’s Guild. In 2013, the group sued Google for digitizing millions of books, claiming it to be a clear violation of copyright laws. They were asking for $750 for every book that was digitized, coming out to a grand total of $3 billion. The judge ruled in favor of Google, claiming the project directly benefits the whole of society while still respecting the works involved.

Google has been doing great work in bringing information to the masses, but clearly more respect needs to be shown to the authorial wishes of those whose content they are disseminating.  

Featured image courtesy of