I think it's pretty certain they will have flaws. The developers will be concentrating on software for them to work well, not on security checks. This happens in all sorts of other fields (including medical!), I don't see any reason these robots should be different.