I have my VHS-C material scanned (15 years ago) in interlaced mode and I found that for further processing by AI algorithms, I had to decide between a process aimed at processing the interlacing or an algorithm to remove noise and improve details.
Then your question attracted my curiosity.
What are the advantages of keeping interlacing?
In my case, the sawtooths of the interlacing were accentuated with AI because the algorithm understood that they were details to be sharpened. So I had to deinterlace first to avoid that unwanted effect.