State-of-the-art stellar structure and evolution codes fail to adequately describe turbulent convection. For stars with convective envelopes, such as red giants, this leads to an incomplete depiction of the surface layers. As a result, the predicted stellar oscillation frequencies are haunted by systematic errors, the so-called surface effect. Different empirically and theoretically motivated correction relations have been proposed to deal with this issue. In this paper, we compare the performance of these surface correction relations for red giant branch stars. For this purpose, we apply the different surface correction relations in asteroseismic analyses of eclipsing binaries and open clusters. In accordance with previous studies of main-sequence stars, we find that the use of different surface correction relations biases the derived global stellar properties, including stellar age, mass, and distance estimates. We furthermore demonstrate that the different relations lead to the same systematic errors for two different open clusters. Our results overall discourage from the use of surface correction relations that rely on reference stars to calibrate free parameters. Due to the demonstrated systematic biasing of the results, the use of appropriate surface correction relations is imperative to any asteroseismic analysis of red giants. Accurate mass, age, and distance estimates for red giants are fundamental when addressing questions that deal with the chemo-dynamical evolution of the Milky Way galaxy. In this way, our results also have implications for fields, such as galactic archaeology, that draw on findings from stellar physics.