Studies in patients and animal models of autoimmune diseases describe different roles for NK T and iNK T cell subsets. For example, CD4+ iNK T cells prevent T1D in NOD mice, whereas iNK T17 cells aggravate the disease. In MS patients under remission, CD4+ iNK T cells secrete large amounts of IL-4, suggesting a beneficial role of these cells. In contrast, CD4+ iNK T cells infiltrate lesions in psoriasis and atherosclerosis, and might be pathogenic. All these data suggest that protection or exacerbation of autoimmune diseases by iNK T cells may be due to disequilibrium between the different subsets (Fig. 3). As highlighted in previous reviews [129,130], most studies used methods that do not identify iNK T cells clearly (e.g. Vα24 PCR, TCR+ NK1·1+ or CD56+ CD3+ staining). CD8+ iNK T cells, representing 20% of human iNK T cells in blood, were rarely analysed and most of the studies focused on CD4+ or DN cells. Similarly, the only cytokines produced by NK T cells were IL-4 and IFN-γ, and only a few studies explored the secretion of IL-2, IL-5, IL-13, IL-17, granulocyte–macrophage colony-stimulating factor (GM-CSF), TGF-β or chemokines. It is important to note that iNK T cell number is higher in mice than in humans, whereas type II NK T  and MAIT cells  are more abundant in humans than in mice. CD1a,b,c-restricted T cells are present in humans, but not mice, due to deletion of CD1 genes in mice, which suggests that CD1d-restricted NK T cells might compensate for these cell populations. Humanized mice expressing these molecules have been generated , and it would be interesting to cross them into genetic backgrounds susceptible to autoimmune disease. Future human studies should focus on less characterized innate T cells. Similarly, the role of self-ligands, cytokine environment and accessory molecules (e.g. NKG2D) required for NK T cell activation should be investigated further [130,134,135]. More extensive research must be performed on specific tissues. For example, the characterization of NK T cells in the pancreas of type 1 diabetic patients remains to be investigated. New mouse models of autoimmune diseases would be useful, in particular, to understand apparent discrepancies in the role of NK T cells between mouse models and human diseases . For example, in mouse models of RA, iNK T cells have a deleterious role, whereas in human RA, iNK T cells seem to exhibit a protective role. It would be interesting to analyse the behaviour of human iNK T cells in humanized mice reconstituted with human stem cells.