Hi @kilbride and @carlgbush. I perform CRISPR/Cas9 gene editing in human pluripotent stem cells, so I am familiar with these topics. Yes, self damage and apoptosis (termination of the whole cell, "programmed cell death") can occur. Pretty much nothing in biology is 100%, so the desired editing events, particularly transmission of the gene of interest, is a fairly rare event. But, there are many things to consider when designing your gene targeting strategy. My understanding is that gene drive in its current form is more used to remove a gene from a population. Removing a gene is a much easier process than making a specific change because you are only relying on the repair machinery of the cell to accidentally repair the double stranded DNA break with an insertion or deletion (termed indel) of one or a few additional/less DNA bases, thus causing a frameshift mutation and essentially removing a gene product.

I would think to actually integrate the "homing endonuclease" (really a Cas9 and a few guide RNAs targeting the region of the genome that you want) directly into the gene that you want to remove would be most efficient, particularly if you were to add a stop codon just downstream of your homing endonuclease. So if the repair process is through homologous recombination, you will end up removing the gene that you wanted to target. However, if the repair process is non homologous end joining (NHEJ), you'll still remove the targeted gene of interest because you'll hopefully have created an indel, rendering the gene product functionally useless.