The problem is that no one has shown FI is necessary to produce the biodiversity we see today. FI can’t be measured quantitatively, so there is no way of determining if it has increased or decreased after a mutation.
If you think I am wrong, then I would suggest a simple experiment. Google for a random DNA sequence generator and start with something smallish, like a 5k base pair random sequence. Measure the functional information in that sequence. Next, make 10 random mutations to that sequence and remeasure the FI and see if it has increased or decreased. I would be interested in seeing the actual numbers.
From what I have read, natural genetic engineering is just random mutations (with respect to fitness) and natural selection.