Will super-smart AI be attacking us anytime soon?

What practical AI attacks exist today? More than zero is the answer and theyre getting better.