In 2000, Shapiro et al. provided compelling "proof of principle" data showing that the transplantation of human islets, purified from cadaveric material, could restore severely diabetic, Type 1 patients to insulin independence. This demonstration prompted renewed efforts to find an alternative and sustainable source of surrogate islet cells for cell therapy. Experiments involving adult ductal and liver "stem" cells, or embryonic stem cells, are prominent amongst these endeavors and are reviewed in this article. Whilst there are many published claims to success in converting ES cells into insulin secreting, glucose responsive cells, all require careful reinterpretation in the light of findings that cells can adsorb insulin present in growth media. It is likely that work with adult cells is less prone to this potential artifact and significant progress has been made in producing insulin-secreting cells. Assessment of in vivo function in the surrogate cells is most frequently made using cell transplantation into toxin-induced, diabetic mice, but this model is rarely used to maximal advantage. In many cases, it remains unclear whether reductions in the hyperglycemia result from insulin secretion from the transplanted cells or are due to recovery of endogenous islet function. In this latter context, experiments are reviewed where endogenous stimulation of recovery is engendered even by irradiated donor cells.