We investigate the feasibility of industrial-scale geological carbon sequestration as an effective climate change mitigation technology. We do so by performing numerical simulations of coupled multiphase flow and geomechanics on a representative field-scale geological model in the Gulf of Mexico that could become a typical CO2 storage site. We use our representative model to investigate two potential hazards associated with large-scale CO2 storage: (1) The potential that CO2 might migrate to shallower formations and (2) The potential that CO2 injection might induce seismicity due to destabilization of pre-existing faults. At our study location, we find that extensive distribution of clay minerals in fault zones in combination with large fault throws result in widespread occurrence of low-permeability faults that prevent migration of CO2 outside of the injection interval. Low-permeability faults act as flow barriers and can result in reservoir compartmentalization and pore pressure build-up. For our field-scale geological model, we find that the risk of induced seismicity is controlled by the choice of the injection location with respect to pre-existing low-permeability faults. Our model results indicate that industrial-scale CO2 storage in the Gulf of Mexico is feasible, but that a priori reservoir characterization and the choice of injection location are crucial to mitigate the potential occurrence of induced seismicity and CO2 migration to shallower formations.