Unless the AI actually intended to kill humans, I don't think it would count as "going rogue" -- that seems more like an accident, or possibly negligence. Note

Unless the AI actually intended to kill humans, I don't think it would count as "going rogue" -- that seems more like an accident, or possibly negligence. Note that our current law draws major distinctions between killing someone by accident, killing someone through negligence, and killing someone with intent.