In February 2023, the editors of Clarkesworld, a prominent online science fiction and fantasy magazine, found themselves in an unusual predicament: they were receiving so many submissions generated by artificial intelligence that they had to temporarily close their doors to new manuscripts. Neil Clarke, the magazine's founder and editor, first noticed something amiss in late 2022, around the time OpenAI released ChatGPT. By February, the volume of AI-generated stories had become so overwhelming that Clarke made the decision to halt submissions entirely.
"We had reached the point where we were on track to receive as many generated submissions as legitimate ones," Clarke told New York Magazine. The magazine, which pays 12 cents per word for accepted stories, had become a target for individuals hoping to earn quick money by submitting AI-generated text. The influx of low-quality, machine-produced content threatened to drown out the work of genuine writers.
The problem is not unique to Clarkesworld. Across the internet, AI-generated content—often referred to as "AI slop"—has proliferated, from fake product reviews and academic papers to entire websites designed to attract ad revenue. Amazon has also seen a surge in AI-written books, many of which are incoherent and pointless. For small publishers like Clarkesworld, the trend poses an existential threat, potentially overwhelming the market for human creativity and original ideas.
Building a Spam Filter
In the months following the shutdown, Clarke, with the help of volunteers, developed what he described as a "very rude rudimentary spam filter" to screen out AI-generated submissions. The filter has been effective in holding the tide at bay, but Clarke remains cautious. In a February 2023 blog post titled "A Concerning Trend," he wrote, "It's clear that business as usual won't be sustainable. If the field can't find a way to address this situation, things will begin to break." He also dismissed claims that AI would spell the death of short fiction, calling such predictions "nonsense," but acknowledged that the situation would complicate things.
Clarke attributes some of the influx to the influencer economy, where social media personalities promote get-rich-quick schemes involving AI tools. "People waving a bunch of money on YouTube or TikTok videos and saying, 'Oh, you can make money with ChatGPT by doing this,'" he explained. This has encouraged a wave of opportunists to flood submission systems with AI-generated content, hoping to cash in on the magazine's payment rates.
The broader implications for the publishing industry are significant. As AI language models become more accessible and sophisticated, the challenge of distinguishing human-written work from machine-generated text will likely intensify. For now, Clarkesworld has managed to keep its submission system functional, but the long-term sustainability of such measures remains uncertain. Clarke's experience serves as a cautionary tale for other publications and platforms grappling with the rise of AI-generated content.
While the immediate crisis has been averted, the underlying issue persists. The magazine's struggle highlights the need for industry-wide solutions, whether through better detection tools, changes in submission policies, or broader discussions about the role of AI in creative fields. As Clarke noted, the current approach is merely "holding things at bay," and without a more permanent fix, the pressure on human creators is likely to grow.
