According to the Edmonton protocol, human islet transplantation can result in insulin independency for periods longer than 3 years. However, this therapy for type 1 diabetes is limited by the scarcity of cadaveric donors. Owing to the ability of embryonic stem cells to expand in vitro and differentiate into a variety of cell types, research has focused on ways to manipulate these cells to overcome this problem. It has been demonstrated that mouse embryonic stem cells can differentiate into insulin-containing cells, restoring normoglycaemia in diabetic mice. To this end, mouse embryonic stem cells were transfected with a DNA construct that provides resistance to neomycin under the control of the regulatory regions of the human insulin gene. However, this protocol has a very low efficiency, needing improvements for this technology to be transferred to human stem cells. Optimum protocols will be instrumental in the production of an unlimited source of cells that synthesise, store and release insulin in a physiological manner. The review focuses on the alternative source of tissue offered by embryonic stem cells for regenerative medicine in diabetes and some key points that should be considered in order for a definitive protocol for in vitro differentiation to be established.