[en] We present empirical constraints on the influence of stellar bars on the fueling of active galactic nuclei (AGNs) out to z = 0.84 using a sample of X-ray-selected AGNs hosted in luminous non-interacting face-on and moderately inclined disk galaxies from the Chandra COSMOS survey. Using high-resolution Hubble Space Telescope imaging to identify bars, we find that the fraction of barred active galaxies displays a similar behavior as that of inactive spirals, declining with redshift from 71% at $z\sim <!--\; ???\; -->0.3$, to 35% at $z\sim <!--\; ???\; -->0.8$. With active galaxies being typically massive, we compare them against a mass-matched sample of inactive spirals and show that, while at face value the AGN bar fraction is slightly higher at all redshifts, we cannot rule out that the bar fractions of active and inactive galaxies are the same. The presence of a bar has no influence on the AGN strength, with barred and unbarred active galaxies showing equivalent X-ray luminosity distributions. From our results, we conclude that the occurrence and the efficiency of the fueling process is independent of the large scale structure of a galaxy. The role of bars, if any, may be restricted to providing the suitable conditions for black hole fueling to occur, i.e., bring a fresh supply of gas to the central 100 pc. At the high-redshift end, we find that roughly 60% of active disk galaxies are unbarred. We speculate this to be related with the known dynamical state of disks at higher redshifts—more gas-rich and prone to instabilities than local spirals—which could also lead to gas inflows without the need of bars.