@HarryGindi I am not sure I am correct anymore since I do not seem to use the hypothesis that the ideal I is regular, so
either the authors don't need the hypothesis or I screwed up.
I use Proposition 2.3 of https://core.ac.uk/download/pdf/82323187.pdf . Call the map R/(a_1^m,..,a_n^m) -> R/I^m for
f_m and the map of inverse systems for f. Then according to the proposition, f is an isomorphism iff for each m,
I can find a s > m and a map g:R/I^s -> R/(a_1^m,...,a_n^m) satisfying the requirements of the proposition.