A teratoma is a benign tumor containing a mixture of differentiated tissues and organotypic derivatives of the three germ layers, while a teratocarcinoma also contains embryonal carcinoma cells (EC cells). Experimental teratomas and teratocarcinomas have been derived from early mammalian embryos transplanted into the adult animal (ectopic sites). In the rat, the pluripotency of the transplanted epiblast was demonstrated and a quantifiable restriction of developmental potential persisted after subsequent transplantation of chemically defined cultivated postimplantation embryos. The rat is nonpermissive for teratocarcinoma development and rat pluripotent cell lines have been established only recently. Transplantation of mouse embryos, epiblast, or embryonic stem cells (mESCs) gave rise to teratocarcinomas. The pluripotency of reprogrammed human cells has been tested by a 'gold standard' trilaminar teratoma assay in immunocompromised mice in vivo. Human pluripotent stem cells proposed for use in regenerative medicine such as human embryonic stem cell (hESC), human nuclear-transfer/therapeutic cloning embryonic stem cell (NT-ESC), or human induced pluripotent stem cell (hiPSC) lines, once differentiated in vitro to the desired cell type, should be again tested in a long-term animal teratoma assay to exclude their malignancy. Such an approach led to a recently implemented human therapy with retinal pigmented epithelium. For greater biosafety, the teratoma assay should be standardized and complemented by assessments of mutations/epimutations, RNA/protein expression, and possible immunogenicity of autologous pluripotent cells. Furthermore, the standardized teratoma assay should be directed more to the assessment of EC/malignant cell features than of differentiated tissues, especially after a unique case of human therapy with neural stem cells was found to lead to malignancy. For further resources related to this article, please visit the WIREs website.