Theoretical predictions have suggested that molecular motion at interfaces-which influences processes including heterogeneous catalysis, (bio)chemical sensing, lubrication and adhesion, and nanomaterial self-assembly-may be dominated by hypothetical "hops" through the adjacent liquid phase, where a diffusing molecule readsorbs after a given hop according to a probabilistic "sticking coefficient." Here, we use three-dimensional (3D) single-molecule tracking to explicitly visualize this process for human serum albumin at solid-liquid interfaces that exert varying electrostatic interactions on the biomacromolecule. Following desorption from the interface, a molecule experiences multiple unproductive surface encounters before readsorption. An average of approximately seven surface collisions is required for the repulsive surfaces, decreasing to approximately two and a half for surfaces that are more attractive. The hops themselves are also influenced by long-range interactions, with increased electrostatic repulsion causing hops of longer duration and distance. These findings explicitly demonstrate that interfacial diffusion is dominated by biased 3D Brownian motion involving bulk-surface coupling and that it can be controlled by influencing short-and long-range adsorbate-surface interactions. DOI: 10.1103/PhysRevLett.119.268001 Molecular transport in fluid phases is understood in terms of the Brownian motion of individual molecules and particles, and can, therefore, be predicted and controlled by parameters including the hydrodynamic radius and fluid viscosity. In contrast, the analogous behavior at the interfaces remains poorly understood despite its fundamental interest and technological relevance; i.e., the dynamics of macromolecules at solid-liquid interfaces underlie many applications including chemical sensing, catalysis, lubrication, and adhesion [1][2][3][4][5][6]. While interfacial diffusion is nominally two dimensional (2D) and conventionally described in terms of 2D Brownian motion, longstanding theoretical models [7][8][9][10][11][12][13][14][15][16] have predicted that interfacial mass transport could actually be dominated by "flights" through an adjacent liquid phase, which would dramatically alter the nature of interfacial molecular motion; an understanding of this process is necessary in order to rationally control mass transport at surfaces. Recent experimental results indirectly support these predictions by measuring the 2D projection of trajectories for atoms, molecules, polymers, and nanoparticles, in thin films, at solid-liquid interface, and on lipid bilayers, which can be represented as an intermittent process with periods of apparent immobility alternating with long flights comprising a heavy-tailed distribution [17][18][19][20][21][22][23][24][25][26]. However, the evidence for the presence of three-dimensional (3D) hops remains indirect, and critical aspects of the proposed "hopping" process remain a mystery. For example, theoretical models represent a flight as a series of hops (i.e., ...