I haven't read from them, but as a general thing, GoF research is very directional, so in practice the experiments will pressure a virus to gain a function, but the rest of the functions are usually impacted to a large extent due to the lack of evolutionary pressure.

So you may do selection on virions that target better certain receptors in certain human cells to infect them, and that's useful to know of possible evolutions of a wild virus, but in parallel it might be losing environmental resistance (temperature range, UV light), or maybe damage the expression of some vital protein, or become too pathogenic and die along with the host.

By all means, one could try to perform GoF in live humans to ensure there's no LoF, but that limits enourmously the speed of the research, plus it's usually forbidden to experiment in humans these days.