It is more intolerable with humans but animals should't be subject of either (animal cruelty).
Yes, that is what I am saying. In your example, of course if there were tests you could have done to prevent the outcome you're at fault. But that aside, the outcome is the same but human editing of a gene means you are causing a specific change in a complicated code bypassing natural adaptations and rejections that might be needed for a trait to be amplified. So you're responsible for the outcome even if there was no testing you could have done to ensure it won't cause cancer on some consumers. You are responsible for any deviations caused by your editing,good or bad. With selective breeding you are forcing breeding between animals that would have bred anyways. It is not impossible for the natural course of things to force breeding between those same pairs of animals. The outcome is the same but since you did not do something that could not have happened anyways in the natural course of things you are not responsible for all deviations.
What I meant though was more like deformations and other flaws that would be cruel to the animal. For example selective breeding of dogs has resulted in some breeds where the brees itself is illegal. Some breeds have skulls that become torturesome and painful once the brain grows to a full size. These breeds would die off if it were not for humans that insist on breeding them.
Mistakes are inevitable and az such the question of "is even one mistake acceptable?" must be asked to which I say no it is not. Trial and error are not acceptable because one error is not acceptable. Ends don't justify means.
Maybe someday computing will be powerful enough to model how editing of genetic material will affect the grown animal (100% accounting of unintended consequences).