That's a very important and basic observation.

You missed another example : cancer. Cankerous cells are much better at replicating themselves than normal cells are. Pluricellular organisms have a multitude of systems to keep their component cells in check, yet they still fail at it from time to time. Biology has had billions of years of evolution to fine tune how it enforces cooperation within larger organisms. Can we do better, especially as the components we're considering at our scale may be as complex and clever to us as cells are to an organism? (Meaning we may not have a comparative advantage even though we're subtler than evolution).

So aside from asking what we can do next, I'd like to add : "Can we do something next?" In order to enforce a system within which you won't observe such an effect, you might need to be larger, have more resources than the sum of all you're trying to steer. Otherwise, some part of that system will eventually take over. And even then, chance events may always remain beyond your capacity to control.