arXiv Bans AI-Generated "Slop" with Temporary Submission Suspension

What happened

arXiv, a widely used preprint server for scientific papers, announced a significant policy change on May 15, 2026. The platform will now enforce a one-year ban on submissions from authors whose work is identified as AI-generated and contains substantial errors or fabrications. This move directly addresses the escalating problem of AI-generated "slop" and hallucinations that can compromise the integrity of scientific discourse.

What changed

The new policy, communicated by an arXiv moderator via social media channels, aims to curb the increasing influx of low-quality, AI-generated content that strains the peer-review system. Historically, arXiv has relied on human moderators and community flagging to identify and manage problematic submissions. However, this new, direct punitive measure is designed to actively deter the creation and submission of such material. While arXiv has not detailed the precise criteria or specific thresholds for identifying AI-generated "slop" or triggering a ban, the message is clear: submissions found to be significantly fabricated or hallucinated due to AI generation will face disciplinary action. This represents a notable shift from merely rejecting papers to imposing penalties on repeat offenders within the scientific community.

• Authors submitting AI-generated content with hallucinations will face a one-year submission ban.
• The policy specifically targets the proliferation of "slop" and misinformation within scientific preprints.
• Direct disciplinary action is now in place for problematic AI-generated submissions.

What we measured

To understand the potential impact of this policy, we examined arXiv's submission guidelines and moderator announcements. We also reviewed recent discussions within the scientific community regarding AI's role in research. Our analysis focused on the practical implications for researchers and the potential challenges in detecting and enforcing such a ban. We considered the types of AI tools commonly used for research assistance, such as large language models (LLMs) like GPT-4 and Claude 3, and how their outputs might be flagged. For instance, a paper generated entirely by an LLM without critical human review could produce factual inaccuracies or nonsensical statements, fitting the description of "slop."

Why it matters for agencies

This development underscores a broader trend of platforms implementing stricter controls over AI-generated content, particularly when accuracy, credibility, and workflow efficiency are at stake. For marketing and content agencies, this serves as a crucial warning. It highlights the need for extreme caution when using AI for content creation, especially for client-facing materials, research-based content, or any output that requires factual accuracy. While arXiv's immediate concern is scientific integrity, the underlying principle of penalizing AI-generated inaccuracies could easily extend to other professional domains. Agencies that utilize AI for drafting research summaries, generating technical documentation, or even producing SEO content must ensure rigorous human oversight and thorough fact-checking. Failure to do so could lead to repercussions, much like how tools such as Jasper AI or Copy.ai require careful human editing to ensure quality and accuracy.

How arXiv's policy compares to other platforms

arXiv's decision to ban submitters is a strong stance compared to many other platforms. For example, Google Search has been refining its policies around AI-generated content, focusing on helpfulness and originality rather than outright bans, as outlined in their [helpful content system updates](https://developers.google.com/search/docs/fundamentals/policies/ai-generated-content). Academic publishers are also grappling with AI, with many requiring authors to disclose AI use in their submissions. However, arXiv's direct punitive measure for detected "slop" is a more aggressive approach, aiming to deter problematic submissions at the source. This differs from simply requesting disclosure or flagging content, suggesting a higher level of concern about the potential for AI to degrade the quality of scientific preprints.

Pros and Cons of arXiv's New Policy

Pros:

• Improved Quality Control: The ban acts as a strong deterrent against the submission of low-quality, AI-generated content, potentially improving the overall signal-to-noise ratio on the platform.
• Reduced Moderator Burden: By discouraging the submission of "slop," the policy could lessen the workload for human moderators who spend time sifting through and rejecting inaccurate papers.
• Upholds Scientific Integrity: The measure signals a commitment to maintaining the credibility of preprints as a reliable source of early-stage research findings.

Cons:

• Detection Challenges: Accurately identifying AI-generated content and distinguishing "slop" from genuine, albeit flawed, human research can be technically challenging and prone to error.
• Potential for False Positives: Authors who use AI as a legitimate research assistant might be unfairly penalized if their work is misidentified.
• Appeals Process Complexity: Establishing a fair and effective appeals process for banned authors will be crucial but potentially complex to manage.

What to watch next

The implementation and enforcement of arXiv's new policy will be closely watched. Key areas of interest include the specific detection methods employed, the clarity of the criteria for identifying "slop," and the robustness of the appeals process for authors who believe they have been wrongly banned. We will be monitoring whether this policy effectively improves the quality of submissions on arXiv and if other academic preprint servers or journals adopt similar stringent measures in response to the growing challenges posed by AI in scientific research. The long-term impact on scientific communication and the ethical use of AI in research will be significant.

---

Source: Ars Technica: Preprint server arXiv will ban submitters of AI-generated hallucinations Source: Google Search Central: AI-generated content

Frequently asked questions

What is arXiv?

arXiv is a free, open-access online repository of electronic preprints of scientific research papers. It is widely used by researchers in fields like physics, mathematics, computer science, and economics to share their work before formal peer review and publication.

What is considered "slop" by arXiv?

"Slop" in this context refers to AI-generated content that contains significant errors, fabrications, or hallucinations. This means the AI has produced inaccurate information or nonsensical statements that undermine the credibility of the research.

How will arXiv detect AI-generated content?

arXiv has not publicly detailed its detection methods. However, it is likely to involve a combination of AI detection tools, human moderation, and community flagging to identify suspicious submissions.

What happens if an author is banned?

If an author's submission is identified as AI-generated "slop," they will face a one-year ban from submitting new papers to arXiv.

Can a banned author appeal the decision?

While not explicitly detailed, it is expected that arXiv will have an appeals process for authors who believe their work was wrongly flagged. The specifics of this process are yet to be clarified.

How does this policy affect legitimate AI tool usage?

The policy targets "slop" and fabrications, not the use of AI as a research assistant. Authors who use AI tools responsibly for tasks like grammar checking or literature review, and who ensure the accuracy of the final output, should not be affected. The key is human oversight and factual accuracy.

Bottom line

arXiv's decision to ban submitters of AI-generated "slop" marks a critical moment in the ongoing debate about AI in academia. By implementing a one-year submission suspension, arXiv is taking a firm stance against content that compromises scientific integrity. While the exact detection methods and appeals process remain to be seen, this policy highlights the growing need for platforms to address the challenges posed by AI-generated misinformation. For researchers, it underscores the importance of rigorous fact-checking and human oversight when using AI tools. This proactive measure aims to preserve the quality and trustworthiness of preprints, ensuring arXiv remains a valuable resource for early-stage scientific communication.