Relying just on h-index is bad in the first places. I habe colleagues in experimental physics who are on papers with tens of coauthors, working only on a small bit that enabled multiple experiments and have an amazing hindex. you cannot compare that to a theoretically working niche field. In the end it is about wrong incentives in science. See also this article on rising self citations [0] . Having said that h-index is a good KPI to track for myself.
I know lots of people who are authors on papers they've never even looked at. (They implicitly agreed to be authors and know that there's some such papers, they just didn't even read the tiles)
Just out of curiosity, do AI practitioners utilize AI to optimize citations? If so, I'd expect papers detailing how to go about it to get massive citations. The appropriate venue might be the Journal of Data Science, Informetrics, and Citation Studies [0], and I didn't see any such articles there yet.
One problem in niche areas is that sometimes there are only so many people working on a particular problem and only so many different papers/authors to cite.