It largely would, depending on your definition of ‘double stranded’. If the great majority of the RNA formed an un-gapped, continuous double helix, then it would be MOSTLY GC AU pairs, so the rule would apply.
But RNA helices are tolerant of GU pairs… so the amount of GU pairing would throw the GC AU ratios off
If we’re not talking GENOMES, but ‘any functional double-stranded RNA’ then tRNA would’ve given Chargaff nightmares! Look at this image
Note FIRST that there are large bits (the ‘loops’ primarily) that are NOT double-stranded, so these can throw off the ratios bc they would be ‘un-partnered’
SECOND, note that tRNAs contain a LOT of modified bases (non A, G, C, U). These are synthesized as AGCU, but modified AFTER synthesis to be weirdos. So Chargaff would’ve had more than 4 ‘players’ if he was studying tRNA.
Chargaff’s rule is a straightforward consequence of two things:
G pairs with C, A pairs with T EXCLUSIVELY in DNA
Genomic DNA is essentially exclusively double stranded, so there would be no regions with exceptions in it (not looking at you, telomeres)