Social choice is concerned with designing and evaluating methods for collective decision making. However, it somewhat neglects computational issues: the problem is generally considered to be solved when the existence (or the non-existence) of a procedure meeting some requirements has been shown; more precisely, knowing that the procedure can be computed is generally enough; now, how hard this computation is, and how the procedure should be implemented, have deserved less attention in the social choice community. This is where AI (and operations research, and more generally computer science) comes into play. As often when bringing together two traditions, AI brings new methods for addressing old issues, but also raises many new issues. This talk considers several of these issues, some of which are on-going and some others are for further research.