The current obsession with citation counts and publication volume creates a perfect environment for generative AI misconduct. When tenure depends on a sheer numbers game, researchers face immense pressure to churn out papers, often using AI to bypass rigorous thought. To fix this, institutions must shift the focus from how much a scientist produces to how much they actually contribute to their field.
One practical step involves replacing broad metrics with qualitative peer assessments. Instead of counting papers, committees should evaluate the impact of a researcher's most significant works. This might mean requiring a 'narrative dossier' where candidates explain the logic, challenges, and long-term influence of their core findings. If a researcher produces three groundbreaking papers instead of thirty mediocre ones, their portfolio should reflect that weight.
Data transparency can also act as a guardrail. Universities could mandate the disclosure of AI usage in all submitted manuscripts and require raw data sets to be readily available for verification. When the prize shifts from volume to reproducibility and intellectual originality, the incentive to use AI as a shortcut to publication disappears. We need to reward the slow, messy process of real discovery rather than the fast, clean output of a machine.
