Ample evidence supports the significance of the high‐latitude ionospheric contribution to magnetospheric plasma. Assuming flux conservation along a flux tube, the upward field‐aligned ion flows observed in the magnetosphere require high‐latitude ionospheric field‐aligned ion upflows of the order of 108 to 109 cm−2 s−1. Since radar and satellite observations of high‐latitude F region flows at times exceed this flux requirement by an order of magnitude, the thermal ionospheric upflows are not simply the ionospheric response to a magnetospheric flux requirement. Several ionospheric ion upflow mechanisms have been proposed, but simulations based on fluid theory do not reproduce all the observed features of ionospheric ion upflows. Certain asymmetries in the statistical morphology of high‐latitude F region ion upflows suggest that the ion upflows may be generated by ion‐neutral frictional heating. We developed a single‐component (O+), time‐dependent gyro‐kinetic model of the high‐latitude F region response to frictional heating in which the neutral exobase is a discontinuous boundary between fully collisional and collisionless plasmas. The concept of a discontinuous neutral exobase and the assumption of a constant and uniform polarization electric field reduce the ion guiding center motion in the frame of a convecting flux tube to simple one‐dimensional ballistic trajectories. We thus are able to analytically calculate a time and height‐dependent ion velocity distribution function, from which we can compute the ion density, parallel velocity, parallel and perpendicular temperature, and parallel flux. Using our model, we simulated the response of a convecting flux tube between 500 km and 2500 km to various frictional heating inputs; the results were both qualitatively and quantitatively different from fluid model results, which may indicate an inadequacy of the fluid theory approach. The gyro‐kinetic frictional heating model responses to the various simulations were qualitatively similar: (1) initial perturbations of all the modeled parameters propagated rapidly up the flux tube, (2) transient values of the ion parallel velocity, temperature, and flux exceeded 3 km s−1, 2 × 104 K, and 109 cm−2 s−1. respectively, (3) a second transient regime developed wherein the parallel temperature drops to very low values (a few hundred Kelvins), and (4) well after heating ceased, large parallel temperatures and large downward parallel velocities and fluxes developed as the flux tube slowly returned to diffusive equilibrium. The ion velocity distributions during the simulation are often non‐Maxwellian and are sometimes composed of two distinct ion populations.