Thea Cooper's book Breakthrough, about the discovery of insulin, has a bit in it about how McLeod, who ran the lab which funded Banting's work, had a theory that it would be easier to produce insulin from certain types of fish pancreases than from pork and beef pancreases because of how fish pancreases kept the insulin producing cells separate from the other bits and pieces. But it came to nothing. It might be that your source for your question has picked up on that aspect and, as often happens on the internet, concluded, incorrectly, that fish are loaded with insulin.
Fish produce insulin, otherwise they wouldn't live (are there T1 fish!?!) but just like non-T1 humans, it's in the pancreas, which is part of the gut, and no-one eats a whole ungutted fish unless they want to spend a week on the toilet.
Anyway, insulin is a protein, so even if you were to find a way of eating a whole stack of raw fish pancreases, any insulin in them would be broken down and would no longer be insulin long before it reached your bloodstream. That's why there are no insulin pills.
You'd be cooking them anyway, and that would break down the insulin.