It seems the issues brought up can be broken down into two lines from the article

the community needs to figure out how to use A.I. “in a way that will maintain human understanding of the mathematics.”

A.I.- generated papers and proofs that have turned out to be incorrect, and in ways that are difficult for mathematicians to discern.

The latter of which is plaguing software developers already who are spending more time debugging vibe code than it would take to write code themselves. The former is more intriguing though in that it would signal glimpses of super intelligence aka cognitive powers beyond top tier experts in a field. Definitely should encourage us all to solve the control problem although major AI labs across the global have seemingly thrown caution Into the wind to solve the funding / cash flow problem first.