Innate and antigen-specific antiviral immunity are triggered by immunorecognition of viral nucleic acids. The helicase retinoic acid-inducible gene I (RIG-I) (also known as DDX58) is the key sensor of negative strand RNA viruses in the cytosol of cells. RNA containing a triphosphate at the 5'-end was shown to activate RIG-I, but the exact structure of RNA supporting 5'-triphosphate recognition, the requirement of a 5'-triphosphate group, as well as the existence of RNA structures detected by RIG-I in the absence of 5'-triphosphate remain controversial. Here, we revisit the literature on RIG-I and RIG-I ligands. The literature proposes at least six different RIG-I ligands: (i) single strand with a 5'-triphosphate, (ii) double-stranded RNA with a 5'-triphosphate, (iii) 5'-triphosphate single-stranded RNA with A- and U-rich 3'-sequences, (iv) double-stranded RNA of intermediate length (>300 and <2000 bp) without 5'-triphosphate, (v) blunt-end short double-stranded RNA (23-30 bp) without 5'-triphosphate, and (vi) short double-stranded RNA (23-30 bp) with 5'-monophosphate. RIG-I thus seems promiscuous for a variety of different RNA molecules, very similar to the Toll-like receptors, of which 10 family members are sufficient for the safe detection of the microbial cosmos. In the light of these outstanding publications, it seems an unlikely possibility that there is a fundamental shortcoming in the design of all studies. Looking closely, the only issue that comes to mind is the in vitro transcription technique used by all investigators without confirming the identity of RNA products. This technique, together with the different biological systems used, the lack of dose responses and of proper comparison of different published ligands and controls leave us with more questions than answers as to what the exact RIG-I ligand is, if in fact it exists.