Background: The IMRaD (Introduction, Methods, Results, and Discussion) format has dominated scholarly communication for decades, yet the integration of artificial intelligence into research workflows and the evolution of computational methodologies have raised questions about its continued suitability for contemporary knowledge production and dissemination. Objective: This study critically evaluated the alignment between the traditional IMRaD structure and AI-integrated research practices, examining factors associated with researcher satisfaction and preferences for alternative communication frameworks. Methods: A mixed-methods convergent parallel design was employed with 847 researchers from 15 countries across five disciplinary domains (computational sciences, life sciences, social sciences, interdisciplinary research, and digital humanities) between March and September 2024. Data were collected through a validated 67-item questionnaire (Cronbach's α=0.89) and 45 semi-structured interviews. Univariate analyses examined distributions of demographic and practice variables; bivariate analyses employed ANOVA, chi-square tests, and Spearman correlations to explore relationships between discipline, AI tool usage, and format satisfaction; multivariable analyses utilized multiple linear regression to identify predictors of IMRaD dissatisfaction and binary logistic regression to determine factors associated with preference for alternative formats, with statistical significance set at p