It might not be practically possible. Genetic modification has side effects, where typos are introduced in other areas than the target.
This means that the first generation of a hypothetical new gene drive is less likely to survive long. But once the gene drive got established subsequent generations would become more and more normal (as typos were lost by interbreeding and natural selection).
Now imagine that we discover a rogue gene drive introduced in humans after, say, it has affected 1000 individuals. There isn't any way of ethically fixing it in their children. You can't go to a person and tell them that they are not allowed to have children. Nor can you tell them that they need to undergo an experiment that makes their children 100x more likely to die young, for example.
For example, there was a story in the UK press (it may be fictional, but I think it will still serve the argument) that there is a gene which causes extremely strong male-directed sexual attraction in both men and women [0] — all men with it are gay, all women with it are hyperfertile, and the hyperfertility is strong enough that positive evolutionary pressure from the women counterbalances the negative pressure from the men.
I think that this sort of thing could count as a toy model of gene-drive, but do correct me if I’m misusing terms or generally failing to understand.
But, even in this toy example probably based on a journalist failing to understand the science, we can not only genetically modify ourselves, but also we have in-vitro fertilisation so we don’t even need to GM ourselves.
[0] don’t worry, I’m bi, this isn’t a hopeless pearl-clutching story about gay genes