While mutant K-Ras is an important therapeutic target for human cancers, there are still no drugs that directly target it. Recent promising studies emphasize the significance of dynamics data to selectively target its active/inactive states. However, despite tremendous information on K-Ras, the direction of information flow in the allosteric regulation of its dynamics has not yet been elucidated. Here, we present a novel approach that identifies causality in correlated motions of proteins and apply it to K-Ras dynamics. Specifically, we analyze molecular dynamics simulations data and comprehensively investigate nucleotide-dependent intrinsic K-Ras activity. We show that GTP binding leads to characteristic residue correlations with relatively long decay times by stabilizing K-Ras motions. Furthermore, we identify for the first time driver-follower relationships of correlated motions in the regulation of K-Ras activity. Our results can be utilized for directly targeting mutant K-Ras in future studies.