Clarkesworld, a science fiction magazine, has banned the submission of new stories after receiving hundreds of low-quality AI-generated pieces.

by matthew sparkes

science fiction magazine clarkesworld has stopped submitting stories after receiving an increasing deluge of AI-generated pieces. The magazine’s founding editor, Neil Clarke, says the problem has been created by people promoting surprisingly capable AI language models like ChatGPT as a way to make money publishing fiction, despite poor quality. of AI stories.

“The typewritten submissions we have received are far from publishable quality,” he says. “I’m sure there are some that are less detectable, but most we’ve received have been easy for me to identify.”

Clarke says he has spoken to other magazine publishers currently facing the same problem, though he says they have been reluctant to talk to the press, like him, until the problem has grown to untenable levels.

The magazine has an open submission policy to encourage new writers, but made the decision to close submissions on February 20 after receiving 50 AI-generated story submissions that day. At that point, clarkesworld it had received 700 legitimate submissions since the beginning of the month and 500 machine-generated, and the rate of increase meant that AI-generated stories would soon take over.

Clarke says that the increase in “spam” submissions began in late 2022, at the same time that accessible big-language AI models were released, and has increased month-on-month. “They don’t care about their reputation on the field. That makes it more like malware or credit card fraud, and faces similar challenges. We will try to minimize these cases and they will try to get around them, ”he says.

However, not all paid science fiction magazines have had the same problem. Djibril al-Ayad, editor of the future fire, says he hasn’t noticed an increase in spam submissions, but that may be because it’s a smaller title. “I guess the problem is that if your posts pile up triple in size and all the AI ​​stuff is abysmal but still has to be read, then it becomes kind of a denial of service attack situation,” he says, referring to to the attackers. that take servers offline when sending a large number of data requests.

“My opinion would be not to be particularly concerned about the danger of accidentally publishing AI-generated fiction,” he says, since most of it will be bad. “If it’s not, then well…great.”