Many problems exist in current attempts to develop polycrystalline GaAs as a basis for thin-film solar cells. Some of these problems arise from the direct interaction of carriers, both dark and photo-generated, with grain boundaries. Others are more indirect; e.g., shunting currents due to the grain boundary-enhanced diffusion of contaminating impurities. This paper describes several of these effects, including the influence of system chemistry on grain properties, the correlation of device parameters with grain size, and grain boundary passivation experiments. A review of various approaches to solving the problems confronting the field is given, and an attempt is made to interpret reported observations in terms of existing theoretical models.