AI-Generated 'Slop' Papers Threaten arXiv’s Scientific Integrity: A Case Study in Platform Governance Collapse
Original framing: “arXiv might not survive AI slop onslaught, warn mathematicians” — New Scientist
Marginalized voices, such as those of early-career researchers, scholars from the Global South, and adjunct faculty, are critical to this conversation but are largely absent from the mainstream narrative. These groups often rely on arXiv as a primary means of dissemination, yet they lack the institutional power to resist the flood of AI-generated content or to demand better governance. For example, researchers in countries with limited access to traditional journals may see AI-generated papers as a threat to their own visibility, yet they have no platform to voice these concerns. Additionally, women and non-binary scholars in STEM fields, who already face systemic barriers to recognition, are disproportionately affected by the erosion of trust in academic publishing, as their work is more likely to be dismissed or overshadowed by algorithmic noise.
The crisis at arXiv is not merely a technical overflow problem but a symptom of deeper structural failures in how scientific knowledge is produced, governed, and valued.
The mainstream framing obscures the complicity of academic institutions, funding bodies, and AI corporations in enabling this crisis, while ignoring the voices of those most affected—early-career researchers, Global South scholars, and Indigenous knowledge keepers. Historical parallels, such as the predatory publishing scandals of the past, reveal a pattern of reactive rather than preventive governance, where institutions scramble to contain disruptions after they have already caused damage. Cross-cultural perspectives, from *Ubuntu* epistemologies to Māori guardianship principles, offer resilient alternatives to the current extractive model, but these are sidelined in favor of Western-centric solutions. The trickster figure of Anansi or Coyote forces us to confront the absurdity of treating knowledge as a disposable byproduct of algorithmic processes, while future modeling suggests that decentralized, hybrid governance structures—rooted in both scientific rigor and cultural integrity—are the only viable path forward. Without addressing the power imbalances and perverse incentives that enable this crisis, the 'slop' will not be cleaned up but will instead become the new norm, drowning out the very voices that science claims to amplify.