And they won’t spoiler it for you.
As far as the fake books, I understand the supposed authors are actual writers. I wonder if any of them will feel challenged to write the book described in the Sun-Times article.
There was something similar done at Astounding/Analog when Campbell was editing. Someone wrote reviews of stories by established SF authors that hadn’t been published. Campbell decided to get the authors to write the stories and published them a year after the reviews.
This was my thought on reading that article, as well.
Sometimes authors and film makers end up making a work that they themselves tossed off as an off-the-cuff title. Arthur Conan Doyle did it with the Sherlock Holmes story The Adventure of the Second Stain (1904), which he mentioned in the story The Naval Treaty (1893). The story, when it finally appeared, couldn’t be reconciled with the description of it Doyle (and Watson) gave in 1893, so it doesn’t seem as if Doyle really had such a story even vaguely plotted out a decade before it appeared.
It is amazing that it made it all the way to physical printed publication without someone raising a single eyebrow, but I think it highlights one of the key pitfalls with generative AI; we keep hearing advice to check AI generated output for facts, but the problem is: generative AI, as a product, is most attractive to the category of people who are least likely to check facts - the lazy.
LHOD is obviously right here (and also to say that if the guy is learning his lesson then that is good and a reason for this not to end his career).
But this is also correct! There was more than one person involved here, and while it is always easy to throw the freelancer under a bus, I bet there are a number of people between him and publication who draw salaries based on a job description that will contain phrases like “ensure high standards in our published work” or “fact check” or “review copy for reputational risk” and the like and all of those people are part of the chain of fuck ups here.
The purpose of a system is what it does. This seems to be the output of a system for getting the maximum amount of “content” in front of readers with the minimum of oversight or critical friction.
There have long been complaints both by journalists and readers that there is too much churnalism, as the demand to produce content for an ever-hungry internet leads to a lot of meaningless filler being ground out by overworked journos* who don’t have the time or the incentive to produce anything better. Put AI into that mix - tell people whose livelihood depends on producing eight or ten or twelve stories per day that there is a magic genie who can produce the required amount of pabulum in a fraction of the time - and it is inevitable that something like this is going to happen.
The question is, what happens next? This guy’s caught flack, the Sun-Times is embarrassed, but what is going to fundamentally change? Will writers be given more time to produce a higher standard of work? Will editors stop demanding endless listicles? What part of the incentives that led this to happen is going to be rethought and retuned to reward and value a better way of working?
*Per LHOD’s description of this webpage, this guy used to enjoy writing and take pride in linking to his stories. Then he stopped doing that, then he became the guy who used ChatGPT to shortcut his churnalism. I doubt that’s where he ever planned on ending up, and as much as anything it’s a little bit sad that this is where things led.
Maybe the other people in the chain just pasted the copy into their favourite LLM and asked ‘proofread this for me’
Streamlining workflows for maximum synergy!
Most drinking water eventually ends up in the sewer, so we connected them directly!
One issue is that most newspapers have laid off all their proofreaders years ago as part of budget cuts.