The current 'publish or perish' model creates a dangerous incentive structure. When tenure committees rely heavily on h-indexes and raw paper counts, researchers face immense pressure to use generative AI to churn out mediocre, incremental studies. To fix this, institutions must move toward holistic, dossier-based reviews that treat a single, groundbreaking paper as more valuable than ten derivative ones.
One practical step involves weighting the quality of the research question. Instead of counting citations alone, committees should ask: Does this work solve a long-standing problem or merely add noise to the field? Evaluations should include qualitative statements from trusted peers who can vouch for the rigor of the methodology and the originality of the insight. This makes it harder to pass off AI-generated filler as legitimate scientific progress.
Universities could also implement
