In February 2023, Clarkesworld Magazine, a leading science fiction and fantasy publication, temporarily halted submissions after being overwhelmed by a surge of stories generated by artificial intelligence. Publisher and editor-in-chief Neil Clarke reported receiving 700 legitimate submissions alongside 500 AI-generated ones on the day the intake was stopped, with the ratio rapidly worsening. The magazine’s experience, as reported by NPR, is emblematic of a broader trend: a systemic overload across numerous institutions as generative AI tools become increasingly sophisticated and accessible.
The initial wave of AI submissions to Clarkesworld, according to Clarke in an editorial published in August 2025, stemmed from online “side-hustle” communities promoting the idea of quickly generating and submitting stories for profit. These groups actively encouraged their audiences to utilize AI and target publications like Clarkesworld. Even as this initial surge has stabilized, becoming more sporadic, a second, more concerning pattern has emerged. These newer AI-generated submissions more closely mimic the style and quality of work from established human authors, making them significantly harder to detect.
This influx isn’t limited to the literary world. Newspapers are facing a similar deluge of AI-generated letters to the editor, academic journals are contending with AI-authored research papers and courts are grappling with AI-drafted legal filings. Lawmakers and educators are also experiencing a significant increase in AI-generated correspondence and assignments. The core issue, as highlighted in a recent analysis published by The Conversation, is that systems historically reliant on the time and cognitive effort required for human creation are now being overwhelmed by the sheer volume of AI-produced content.
Institutions have responded in various ways. Some, like Clarkesworld initially, have temporarily closed submissions. Others are employing defensive measures, often involving AI itself. Academic peer reviewers are increasingly using AI tools to identify potentially AI-generated papers, while social media platforms are deploying AI-powered moderation systems. Court systems are utilizing AI to triage the increased volume of filings, and employers are leveraging AI to screen job applications. These responses represent an escalating “arms race,” characterized by rapid iteration and the application of AI to both create and detect generated content.
Still, these arms races are not without unintended consequences. A clogged court system due to frivolous AI-generated cases, or the devaluation of academic merit based on fraudulently submitted AI-written work, represent significant societal harms. The concern, as outlined in The Conversation, is that unchecked AI-enabled fraud could undermine the integrity of institutions society relies upon.
Despite these risks, some applications of AI assistance offer potential benefits. In scientific research, AI can aid in literature reviews, data analysis, and even scientific communication, potentially leveling the playing field for researchers whose primary language isn’t English. Prior to the advent of AI, researchers with greater financial resources could afford human assistance with writing and editing academic papers. AI now provides a more accessible alternative. Similarly, in fiction, while fraudulent submissions are detrimental, some outlets may explore accepting AI-assisted submissions with appropriate disclosure and evaluation criteria.
The key distinction, according to the analysis, lies in the power dynamic. While AI can democratize access to writing and cognitive assistance, it also enables the amplification of misinformation and manipulation. The use of AI to generate multiple letters to the editor and present them as individual opinions – a form of astroturfing – poses a threat to democratic processes. The same technology that empowers citizens to express their views can be exploited by corporate interests to distort public opinion.
Clarkesworld reopened submissions in 2025, stating it had developed an adequate method for distinguishing between human- and AI-written stories, though the long-term effectiveness of this approach remains uncertain. The arms race continues, and the ultimate balance between the benefits and harms of AI remains to be seen. As of February 11, 2026, the magazine continues to accept submissions, but the situation remains fluid.
