So this has little to do with RISC, except the general principle that the instructions that are used more frequently should be implemented to be faster, a principle that has been used by the M1 designers and by any other competent CPU designers.
In this case, ARM has added the polynomial multiplication instruction a few years after Intel, with the same main purpose of accelerating the authenticated encryption with AES. There is little doubt that ARM was inspired by the Intel Westmere new instructions (announced by Intel a few years before the Westmere launch in 2010).
The dedicated CRC32 instruction could have been made much faster, but the designers of the M1 core did not believe that this is worthwhile, because that instruction is not used often.
The polynomial multiplication is used by many more applications, because it can implement CRC computations based on any polynomial, no only that one specified for CRC32, and it can also be used in a great number of other algorithms that are based on the properties of the fields whose elements are polynomials with binary coefficients.
So it made sense to have a better implementation for the polynomial multiplication, which allows greater speeds in many algorithms, including the CRC computation.