Vox Day had a post about Clarkesworld science fiction magazine that had to shut down its submissions due to being overwhelmed by the volume of AI generated novels being submitted. Apparently, they can’t afford the software that could automate the process of identifying and rejecting the AI generated books.
As someone currently writing a science fiction novel this is a remarkable turn of events. Wouldn’t it be something if the artificial authors produce a better product than the organic ones! And I don’t mean more grammatical, but more creative and entertaining. So, if the algorithm blends 60% Heinlein with 20% Dickens and 20% Hemingway would it produce the greatest science fiction story of all time or an abomination not fit for human consumption?
In one sense I’m lucky. I’ve read that ChatGTP is hopelessly woke so my current story about a revolt against an authoritarian Deep State cabal is impossible for ChatGTP to write. If it attempted such a thing it would blow out every circuit in its server farm.
But just thinking about this whole Frankensteinian situation is amusing. I imagine Dr. Morbius from Forbidden Planet describing to a visiting interstellar fiction author the ability of the Krell machine to write sci-fi stories, “You see these row after row of instruments? Each one represents an increase by a factor of ten the number of cybernetic neurons figuring out plot details. It’s the number ten raised literally to the power of infinity of cliffhanger chapter endings and hypercompetent heroes saying something self-serving.”
But then his guest would remind him of monsters from the id, “Yes Morbius, but without the ability reflect reality; without the ability to distinguish male characters from female characters and to allow non-white characters to have normal human flaws it will all quickly devolve into unreadable goo.”
To which Morbius would reply, “Of course, the mindless primitive, why didn’t I see it? Hypersensitive, politically correct story lines are stultifyingly boring. Quick, son, press down this lever and be one hundred million miles away before it blows. I’ve set ChatGTP to overdrive and when that amount of drivel piles up the result will be a literary black hole that will render all of science fiction lifeless within fifty parsecs.”
Well, I wonder how all of this will impact self-publishing at places like Amazon. Will even their servers be able to handle the surge in output from ChatGTP filling up the world’s server farms with romance and porn novels about super-smart super-attractive girl-bosses who save the known universe.
Heaven, help us.
While the thought of more legislation needed to ameliorate this issue is a loathing thought, it could be needed in the near future. It should be required that AI-generated content be labeled as such to inform the consumer. We already have requirements for notification on food if they have been enhanced with hormones, as well as products containing chemicals that can cause harm, so it isn’t too far off where this should be implemented on AI content. Although, I can see where it could lead to a certain point where human generated stories and content become more valuable due to… Read more »
Arthur, since the AI has been created with algorithms that enforce woke talking points it will at least be possible to differentiate one’s work from it by using language that defies things like gender neutral pronouns and other stuff that will annoy leftists. And by golly that’s just what I always do!
Looks like I’ll be shopping by author from now on. This is sad. It will make it difficult for new authors to break in.
No, wait – there will be publishers who will only use human authors. I expect companies like Baen books will be one of them. I’ll just start buying from trusted publishers.
I have to assume there will be a way for the publishers to recognize the AI generated books. They will probably have a “fingerprint” based on word selection and sentence structure. But it could make things difficult until things get figured out.