Things are moving too fast for us humans… some of us, anyway. The proliferation of “artificial intelligence” programs (a term which I continue to view dubiously and speak sarcastically), all of them capable of churning out huge amounts of prose on demand, has us asking about the real origins of various works. A great many writers are anxious about the possibility of being caught in that net, whether or not such an accusation would be accurate.
Have a high-profile, high-dollar case:
A high-profile publishing deal for a debut crime novel has collapsed after doubts emerged over whether artificial intelligence played a role in writing it.
The hotly anticipated manuscript Call Me, I’ll Hide the Body, by Jerry Falade, was withdrawn from sale by its agent despite reportedly receiving an offer for more than $2m (£1.5m) from Minotaur, owned by Macmillan US, as part of a 14-way auction. The plan was to publish it in 2028.
An email sent to publishers by Falade’s agents and seen by the Guardian said: “Jerry Falade produced an incredible manuscript that dazzled us and the top publishing professionals around the world. Unfortunately, we are no longer able to authenticate how the manuscript fully evolved from origin to completion.”
It continued: “During the submission process we vetted the manuscript with Jerry for AI and felt trusting in his reassurance that it was written without utilising AI as a resource in the writing or editorial process. We can no longer substantiate that, and we are withdrawing the book.”
There are several fascinating things about this case:
- The huge amount of money involved (approximately $2,000,000);
- The reaction of agents and editors to the book (“stunningly good” and “appeals to absolutely everyone”);
- The author’s claim that the rejection of the book as possibly AI-generated reflects “racial bias.”
(Yes, Jerry Falade is black. But that’s a topic for another time.)
Now, the “AI detection” software that’s occasionally used to analyze a manuscript does produce false positives; how frequently, I cannot say. Whether an MS suspected of coming from an AI “author” might be worthy even so is for publishers and readers to decide. But publishers are within their rights to say to a human writer that “We don’t think you really wrote this” and reject it. The legal problems involved in copyrighting a novel of doubtful provenance are sufficient reason.
What will follow? Will it be obligatory for human writers to keep copies of old drafts of their novels, dating to before the AI explosion? Considering that virtually all manuscripts today originate on a computer, what would that prove? And what of writers to come, their works not yet complete even in first draft? How would they authenticate their authorship?
This is no common-garden-variety quagmire. AI programs are trained on human works, including both old and new texts. Should your magnum opus be one of them, and some AI incorporates important elements of your style into its own production algorithm, what recourse would you have? Were you unable to produce a paper copy of your book that could be radiocarbon-dated to a date before any AI could have processed it, would there remain a way to prove that you, and not some AI, were the originator?
Things are going to be tough for us indies, in particular. Most of us have no way to “trademark” our prose. AI programs are now common enough that just about anyone who has access to the World Wide Web can employ one. And the “AI detectors” are in use throughout every pocket in the world of fiction.
If there’s a silver lining, it would be that “respected” publishers are now on level ground with indie writers and Johnny-come-lately publishers. We’re all equally vulnerable. Of course, should your trailblazing effort in Vampires Who Teleport or Zombie Romance come under suspicion of having been generated by Claude, Grok, OpenAI, or Gemini, that will be small comfort.
No comments:
Post a Comment