Position: Academic Conferences are Potentially Facing Denominator Gaming Caused by Fully Automated Scientific Agents

A new threat called Agentic Denominator Gaming could exploit AI conferences' stable acceptance rates by flooding submissions with low-quality papers generated by automated agents, inflating the denominator to boost legitimate papers' acceptance odds without intending publication of the spam itself. This systemic vulnerability exposes academic peer review to coordinated attacks that would degrade review quality and increase reviewer burnout while requiring institutional policy reforms beyond technical solutions.

The academic publishing system faces an emerging vulnerability as AI-generated content production scales exponentially. Conference organizers maintain relatively fixed acceptance rates despite growing submission volumes, creating a perverse incentive structure: flooding the system with plausible but low-quality papers increases the denominator, mechanically raising acceptance probabilities for targeted legitimate work without requiring those spam papers to be accepted. This represents a fundamental shift from traditional gaming tactics that seek acceptance of bad papers to a more sophisticated attack exploiting the statistical architecture of peer review itself.

The vulnerability stems from decades of conference policies designed to maintain prestige through stable acceptance rates. As submissions have grown exponentially—particularly in AI conferences—the acceptance rate percentage has remained constant, meaning absolute numbers of acceptances grew proportionally. Automated agent technology now makes generating thousands of superficially plausible papers technically feasible and economically rational for adversaries seeking publication advantages. The threat extends beyond individual gaming to potential industrialization: coordinated agent mills could become commercial operations selling submission services.

The industry impact spans multiple stakeholders. Researchers face degraded signal-to-noise ratios in conference literature and intensified competition. Reviewers experience burnout from processing volume while struggling to identify automated content. Conference organizers confront impossible scaling challenges. This dynamic could accelerate migration toward alternative publication models, AI-powered peer review systems, or invitation-only venues, fragmenting the academic commons. The problem cannot be solved through detection algorithms alone since the goal isn't acceptance of bad papers—institutional reforms around submission policies, review incentives, and acceptance rate mechanisms are necessary structural changes.

• →Agentic Denominator Gaming exploits stable conference acceptance rates by flooding submissions to mathematically increase legitimate papers' publication odds without requiring spam acceptance.
• →Automated AI agents make large-scale paper generation economically feasible, enabling potential commercialized 'agent mills' offering submission services.
• →The threat degrades review quality and accelerates reviewer burnout while evading traditional detection-based defenses.
• →Academic conferences need institutional policy reforms addressing submission fees, acceptance rate mechanisms, and review incentives rather than relying solely on technical detection.
• →This systemic vulnerability could fragment academic publishing, pushing migration toward invitation-only venues or AI-powered review alternatives.