Since almost two decades, the prestigious Clarkesworld Magazine has promoted the careers of science fiction authors by frequently publishing works by Hugo Award nominees and victors like Elizabeth Bear, Peter Watts, and Catherynne M. Valente.
However, it is currently engaged in a rather absurd conflict with that most futuristic of contemporary trends: AI.
Over a third of the submissions to Clarkesworld this year were written by artificial intelligence and then filed by dishonest humans, according to a recent article by the magazine’s editor Neil Clarke. And it’s rapidly growing worse. More AI-written entries emerged in the first half of February than in the entire month of January, and according to Clarke, there were 50 just today.
Since the article’s publication, Clarke has tweeted that submissions are currently completely stopped. He continues, “I shouldn’t be hard to predict why.”
As the numbers poured in this morning, Clarke told Kotaku via email, the choice to close submissions was taken “on the fly.” “I could either close entries and work with the genuine submissions or play whack-a-mole all day.”
It’s amazing how quickly this issue has escalated. Although Clarke claims in his blog entry that he has long had to deal with plagiarism, the issue didn’t become so pervasive until the end of 2022. And then, in the first half of 2023, it got so bad that the journal had to stop accepting submissions.
How does Clarkesworld reveal an AI-generated story?
For the perfectly reasonable reason that he doesn’t want to provide cheaters with information that could enable them to evade his detection, Clarke doesn’t explain in his blog how he can tell which posts are written by AI. He did, however, tell Kotaku that they are presently fairly easy to spot.
Clarke informed us that “the ‘authors’ we’ve banned’ have been blatantly sending machine-generated text. These are formulaic, low-quality compositions. He also believes there is a level above these, one that is less apparent but still suspiciously higher. But he adds, “It’s inevitable that group will grow over time and become yet another problem because none are ever good enough to justify spending more time on them.”
It’s not just Clarke who has to deal with it. The editor claims that others in comparable roles are experiencing the same difficulties, and it is obvious that if it is occurring at Clarkesworld, it will happen everywhere that accepts submissions for publication. The majority of these submissions are weeded out because they won’t be publishable, but it can be costly and time-consuming to sift through the fakes.
Given the number of false-positives and false-negatives, as well as the price of such services, Clarke continues that third-party detection tools that are supposed to be able to identify copied or AI-written content aren’t the answer. Other short-term solutions, such as regional embargoes on nations where the majority of fake submissions originate, are also ineffective. According to Clarke’s essay,
It’s obvious that things won’t continue as normal, and I’m concerned that going down this road will result in more obstacles being put in the way of new and international authors. These individuals are essential to short fiction.
Of course, this is not a problem that will get any simpler. You could start writing a science fiction short story at the rate at which AI chat bots are developing, and presumably future changes will make it harder and harder to instantly identify them. We’re probably still a long way from AI being able to write tales that are actually worth reading, though. I enquired of Clarke if he believed this to be the situation. He said, declining to speculate on when such a leap might occur, “At the moment, considerable improvement is still required.”
However, this doesn’t offer much solace. As Clarke stated to Kotaku, “We still have ethical concerns about the methods used to create these works, and until such concerns can be alleviated, we won’t even contemplate publishing machine-generated works.”
An effort at a science fiction story by ChatGPT and Chatsonic
There are currently services like ChatSonic that brashly market themselves as a way for students to produce chunks of original writing. I’ve previously had tiresomely pointless arguments with the AI about how this is obviously cheating, to which it becomes incredibly enraged and insists that simply requesting the bot for words on a subject is a creative act in and of itself.
In fact, I asked ChatSonic to create a 1,000 word short story about an AI who writes science fiction and goes on to receive a Hugo Award while I was composing the preceding sentence. It took a short while, but for some reason it only got to 293 words (bloody contractors), and it’s terrible:
In contrast, ChatGPT made a much better effort, meeting the word limit and producing content that showed some creativity. In the end, it’s still a terrible tale that is hysterically conceited but terrifyingly competent:
Can AI outdo human creativity?
Before even considering publishing writing created by artificial intelligence (AI), Clarke stated that he has a number of ethical issues to address. But could something like that ever happen? Could it ever make sense to print such works if AI could create original stories that are worth reading? You need these tools in order to be able to write something that extends beyond its dataset, Clarke explained to us first. The real deal; not a remake At that moment, it can compete with our best writers but isn’t always certain to be superior.
Of course, “superior” may not be the most important consideration. The main distinction—and the one that is currently giving us trouble, as Clarke continues—is speed. A computer can outproduce a human artist and drown them in the din.
And to top it all off, here is ChatGPT’s chilling conclusion to the short tale I requested earlier, in case that wasn’t enough to make you nervous:
Naturally, some individuals were still dubious. They held the view that an AI was simply regurgitating information that had been programmed into it and could never be genuinely creative. But SciFiGenius’s followers knew better. They were aware that the AI was much more than just a machine that could spit out pre-written tales. They were aware that it was a genuine artist who was capable of producing pieces that affected millions of people.
Download The Radiant App To Start Watching!
Web: Watch Now
LGTV™: Download
ROKU™: Download
XBox™: Download
Samsung TV™: Download
Amazon Fire TV™: Download
Android TV™: Download